This "patented prediction technology" will be incorporated into the search function, and its advertising and marketing.

Hunch launched in 2009 as a platform for recommending things that it believed its members would like based on what they shared online. It relaunched in 2010 as an Internet personalization service with a taste-graph driven recommendation engine that recommends highly targeted personal recommendations to its users based on 20 quick questions. Hunch is now officially a part of eBay, but will keep its New York-based office, and continue to operate as its own entity.

That's exactly what the team at Hunch has done. The personalized recommendation service launched their browser-based Netflix Predictor today, which uses the company's "taste graph" to help determine what movie you should watch next.

The site does a pretty impressive job of suggesting what to watch. When I connected my Netflix account and started tinkering with the slider-based filters across the top, I found myself genuinely interested in checking out many of the films that popped up. The Web app lets you tweak the results by genre, release date, maturity rating and overall popularity.

The final product is a list of recommendations that is totally separate from Netflix's own native suggestions, as decent as those can be. In some ways, Hunch's are better, especially if you narrow things down using the filters.

You can also watch a trailer of each movie, read a summary and, if your interest is piqued, jump right to Netflix to start streaming, assuming the movie is available on demand.

This one, however, is terrible. There were errors throughout the log-in process (you can't listen to music until you log in either), it required I give the site access to Tweet from my account and the Digster.fm Twitter account has been Tweeting out links all afternoon to websites that posted their raw press release.

Bad music, bad playlist titles, non-helpful one-line descriptions for the playlist. Some goofy "mood indicator" iconography. Terrible. No wonder Universal went to great lengths to keep its name off the site, except on the pages for the legal terms.

A single intern at "taste graph" startup Hunch.com built a little side project that went live today called Hunch.fm. It is much, much better than Digster.

Other sites offer me something else, like meaningful personalization or social discovery. Artist information. Curation I find of interest. This is just half-baked, condescending and pointless. We have the Internet now, Universal. It brings us wonderful things - the bar has been raised. The bar is also higher because you are the biggest of all the music label groups.

Sorry, music industry, you're going to have to do much, much better than this.

Below: I opened the Digster playlist of True Latin Rock and the first thing I see is an album cover of someone with their underpants around their ankles while a band yells the word "Puto!" (Spanish slur for male homosexual prostitute) at me for 2 minutes. Thanks, Curator Angela!

On stage today at ReadWriteWeb's 2WAY Summit, Dixon sat down with our own Marshall Kirkpatrick to talk about how Hunch has built its taste graph and how this sort of recommendation engine may shape the future of a more personalized Web.

By asking just a few simple questions to users, the Hunch website is able to predict with pretty astonishing accuracy how they'll answer other questions. After you "tell Hunch about yourself," the startup's recommendation engine is able to offer suggestions about other things you might like. Kirkpatrick pointed to a recent infographic on Dixon's own blog, detailing the amount of data that the startup is working with: about 500 million people, 200 million items and 30 billion edges. That last figure is key, as it means that Hunch has what Dixon calls a "known preference" for 30 billion items.

Dixon explained a little bit of the technology behind Hunch and behind other recommendation engines. He noted that unlike some systems that rely on natural language processing - on what people say, for example - Hunch is able to glean quite a lot based on who you follow on Twitter and what you like on Facebook. So if you follow Barack Obama, you're more likely to be a liberal. And you're more likely to prefer arugula.

Dixon said that initially Hunch had thought about building its own dataset, but instead has tapped into what's already on the Web, utilizing Facebook and Twitter authorization for example in order to identify some of these tastes.

Hunch Knows What You Like: Privacy Concern?

In light of concerns about Google and Facebook building facial recognition technology, Kirkpatrick asked how (or if ) we can protect people's privacy when it comes to "taste recognition." What are the implications of being able to tell so much about a person - their sexual orientation, their political orientation - just by their answering a few questions or linking their Twitter or Facebook accounts.

Dixon made it clear that despite Hunch's ability to predict users' tastes that the company would never sell that data. "We have never made a data deal," said Dixon. Furthermore, people can only get predictive information about themselves.

What Dixon envisions for Hunch nonetheless is to be "the place you keep your taste profile." That's something that's made a lot easier, in no small part, thanks to the Hunch API. Dixon talked about the possibilities of incorporating the Hunch taste graph into various applications, from better follow suggestions on Twitter to better hotel suggestions on a site like Kayak.

Dixon says that he believes that there seem to be two ways in which personalization will occur in the future - "either through shadowy cookies and things behind the scene" or by things that users control. And we want to be the people doing it."

One of the big reasons Netflix is so successful isn't just that it's cheap and available over the Internet. The service keeps track of what you watch, takes into account your ratings of different content, and then makes suggestions for what else to see according to all of that data. The accuracy of those predictions likely determine your opinion of the service itself.

Today Hunch has partnered up with Samsung and Digitas to launch The Smart Living Room, an "interactive microsite that makes movie watching a deeper social experience."

The Smart Living Room is taking that predictive ability and adding it to television viewing. Here's the explanation from the announcement:

After a viewer answers a series of Hunch personality questions, The Smart Living Room creates a personalized movie recommendation including the genre best suited for that viewer as well as movie title suggestions. The viewer can then create a movie watching event by inviting friends and family through Facebook or email. The site utilizes dynamic CG animations to offer the viewer entertainment and surprises throughout the experience.

If Hunch can bring it's often appallingly accurate predictions to content recommendation, Samsung could have a winner on its hands. Unfortunately, all the site currently recommends is a basic genre, like "Action" or "Comedy", a pretty 3D animation and a couple title recommendations. With a prediction engine like Hunch, we're hoping to see more from this in the near future.

Wikipedia will celebrate its 10th birthday tomorrow and while it's changed the world in incredible ways, it enters its second decade in the same position it began: as a destination website. In an era of web services and applications, Wikipedia could be so much more. Wikipedia as an organization would like it to be more. Unfortunately, it's not well positioned to realize its full potential yet - and the world isn't ready for it, either.

Wikipedia could be a provider of canonical descriptions of all kinds of things, a platform that applications of all sorts could use to populate themselves with rich information on almost any topic. Much like place databases are used by so many location-based technologies today. If you want to build an app that talks about different places around the world, you don't have to create a brand new map of places from scratch. That problem has been solved.

If Wikipedia entries could be used like this, software companies could focus on building fabulous experiences for their users, without having to worry about re-creating descriptions of all thing topics they are discussing.

"There is a fair amount of structured data that is reasonably machine readable," Wikipedia co-founder Jimmy Wales told ReadWriteWeb this week in a press call. "I'm definitely supportive in general of initiatives to take information from Wikipedia and do interesting things with it."

Steven Walling, a former ReadWriteWeb writer now with a fellowship at the Wikimedia Foundation, considers Wikipedia as a platform "an open question" - which must be a wiki nerd's most enthusiastic endorsement. "Our core mission is broad," he told us, "it's not just to create a web based encyclopedia, it's explicitly designed for reuse. Right now, Wikipedia is a destination site, but it's an open question.

It seems that Wikipedians want the site to be used as a service, though it's not clear they really want it enthusiastically.

Right now, Wikipedia isn't being used as a data platform very much - and it appears unlikely that it will be. It's a real loss for everyone.

The Problem With Wikipedia

The Hunch blog this morning announced that the company would no longer use Wikipedia entries to populate the descriptions of the things, like foods, vacations, articles of clothing etc., that it recommended to its users.

The reasons will probably feel familiar to many people:

• Many of the descriptions from Wikipedia were either overly general, overly technical, or very out of context for the specific result on Hunch.


• Search engines tend to penalize sites that have a significant amount of what they consider "non-original" content. Thus, by having many results with Wikipedia-like descriptions, we may have been limiting the visibility of those results in search engines and thus reducing the number of people who could use a search engine to find relevant information on Hunch.

Right: Hunch, describing something now not at all with the words of Wikipedia.

It's possible that the company may have other concerns it's not discussing, regarding its commercial dealings with 3rd parties, but both of those reasons above seem quite valid.
It's a real shame. As Creative Commons licensed content, the Wikipedia corpus seems like something that ought to commoditize high-quality knowledge, allowing software developers to build even more value on top of it. The Hunch team specializes in machine learning and recommendation technology. Why should it have to also recreate descriptions of the whole world it's drawing recommendations from? In theory, that's a solved problem - Wikipedia solved it.

Unfortunately, Wikipedia articles aren't written with this in mind and are often far too technical. They aren't written with re-use in mind, for example, they often don't put the most accessible content at the top.

The search engine problem is a whole other story. The duplicate content punishment clearly hasn't succeeded in weeding spam or low-quality content from search results. It seems like a hold-over from the days when good actors online were creating all their own original content by hand and bad actors were the only ones who used machines to work with data in bulk.

What the duplicate content penalty has done is punish innovative startups like Hunch, who have a lot to add in terms of technology but who would do best to rely on someone else's text data and descriptions of things.

"That's a shame," Walling says, "that someone feels they are being punished for reuse, but we can't change the way search engines work."

"One of the biggest reuse projects right now is Facebook community pages," Walling told us. "They reuse a lot of Wikipedia content. Our business development people worked with them to set that up, but their content isn't indexed by search engines, so that isn't an issue for Facebook."

That is an issue for smaller companies though. This isn't just a loss for Hunch, it's a loss for a whole ecosystem of startups that could serve their users with innovative new technology built on top of a commoditized standard of collaboratively edited descriptions of the world. As users, that means it's a loss for us.

SCVNGR is also switching away from its own proprietary location database. Thanks to its close relationship with Google (SCVNGR is, in part, funded by Google Ventures), the service is the first site to leverage the new Google Places API.]]> As the company's founder and "Chief Ninja" Seth Priebatsch told us yesterday, switching to the Google Places API allowed SCVNGR to quickly scale globally, as it can now rely on Google's extensive location database to power its service. SCVNGR users will now also be able to create challenges and treks - the central gaming elements on the service - at all of these locations. As we noted when Google first announced it, the Places API "could do for check-ins what Google Maps did for maps." As Priebatsch told us, the fact that Google gave his company access to this comprehensive global database with millions of locations made switching to it a no-brainer.

While SCVNGR plans to launch localized versions of its apps in the future, the interface is currently only available in English. Users can create challenges using their own languages, however.

The service, which launched just 20 weeks ago, currently has about half a million users, though it's not clear how many of these are active users. Priebatsch also told us that the company has managed to double the number of enterprise clients since its official launch to over 1000. SCVNGR clearly has a lot of momentum going for it right now and it will be interesting to see how the service develops now that it is going global.

Reddit's goal for this data is to see it used to create a recommendation engine - in particular a system that would highlight some of the niche communities on Reddit that are a great place to find good topical content, but that too few people on the site have discovered. Now that the data is out in the wild, however, any number of analyses can be performed on it - and no one knows what kinds of observations about the relationship between people, web content, voting and news will be discovered. One little account preference opens up a world of opportunities: "allow my data to be used for research purposes."

"It's great to have these kinds of data dumps available for research," says Joel Spolsky, co-founder of the popular StackOverflow network, which makes its user data available in a bulk dump every month, under a Creative Commons license. "We've had several academics analyzing our data dump and learning interesting, measurable, scientifically relevant things about online communities. You never know what's going to come out of it."

Chris Dixon, co-founder of recommendation service Hunch, said the Reddit data and recommendation effort are a "great project." "I think I'll have our devs hack something together using the Hunch API," he said. "We have a blog recommender widget [of our own] coming out soon." Dixon's company is one of the most prominent startups aiming to build a "taste graph" and Hunch already offers recommendations that impress many people, on a wide variety of topics.

Real-world recommendations, profile analysis for increased self-awareness and scientific insights into the nature of online life: those are the kinds of things people are building now with publicly available social network user data.

Nowhere in the world is there more opportunity to develop such insights based on user data than on Facebook. Facebook used to hand over data dumps of its users activities to big companies doing research without communicating that to the users. Now, a much larger company, Facebook is maddeningly unwilling to offer bulk data export for research and analysis. Perhaps in part because people are so upset whenever the release of data is perceived as a violation of online privacy.

That's an easy problem to solve, though, if Reddit is any indication. Just ask users if they want to check one box: "allow my data to be used for research purposes."

Please, Facebook?

I spoke to Caterina Fake to find out how Hunch got started and the progress the company has made in its ambitious mission.

Caterina Fake: It goes back a little bit to when I was still at Yahoo. I decided the most interesting thing going on there was search. I took my social media, user-generated content background and tried to use that as a lens through which to look at search.

I briefly worked on Yahoo! Answers while I was at Yahoo, as well as a bunch of other social search products. And there was a sense that this was one direction that social [search] can take. There was an explosion of sites like Quora and Aardvark. I think the crucial turning point in the way things were evolving was in April of this year, the [Facebook] F8 conference - for us anyway. That was a very significant thing. Facebook announced their "Like" button and an implementation of that across their site and across the web. And using that data in a recommendations. Then Amazon and Google started playing in the same space. There was just a lot going on in that general direction.

I think the most important thing for an entrepreneur is to pick a good problem and find a solution to it. We picked search as a problem to solve and then we just iterated. We launched a decision tree model, based on expert systems (which has been around since the 80s) and we built a whole series of decision trees. The content is mostly user-generated. We launched with about 500 and it was an alternative to search. It was a way of getting a little bit further down the search funnel.

So the kind of thinking behind this was: you could do a search, but you could also get a bit further [because] we will give you results that will give you instantaneously what would otherwise take you several hours of research to do properly.

One of the things that we did, and this was the kind of innovation that we actually did before our launch, was a section of the site that asked questions about you. There are questions about your demographics, your politics, your taste, your beliefs and values - those kinds of things. It's deliberately based on the design of Hot or Not. People were enjoying themselves so much answering the questions that the average user answers over 150 questions. We've got 60 million, now probably approaching 70 million, questions answered from that module. We were able to map that to people's taste. And so we created, as a kind of exhaust from our system, a profile of each user.

We learned that it was an extraordinarily valuable thing to build, because then you could actually skip most of the questions on a decision tree - because we'll take you straight to the answer that we assume is your preference.

RM: So the questions started as just one feature of the product, but it's turned into the foundation of the product?

CF: Exactly. What had been a subordinate feature became the heart of the product. And that is what we are currently focused on.

So, we now have profiles on millions of people and are applying those anywhere on the web. For example, we should be able to use it in Foursquare to determine what restaurant you should go to. Once somebody knows you, or as in this case our system knows, it should be able to become your own personalized navigation ball.

RM: Is there a particular target audience that you are going after?

CF: The way that we're approaching things right now is that we're developing partnerships with various companies. We are approaching verticals, learning as much as we can about each vertical and how that applies to what we call the taste graph. Then using that knowledge through people's recommendations of results. So we're in the throes of tons of tons of learning. We are kind of like a machine learning shop. Our appetite for data is insatiable.

The thing at which Hunch itself is concerned with is the processing of that data. How do you create a taste graph? How do you figure out the affinity between almost infinite numbers of entities. How you do you constrain them, how do you deliver them in real time, how do you assess the adaptability to each particular user from their idiosyncratic taste profile. We're concentrating currently on local and shopping. There are a bunch of things that we can then do.

RM: You mentioned Facebook 'likes' before. Does Hunch hook into Facebook to get that kind of taste data? Or is that a separate thing which you can't touch at this point?

CF: When you're designing a product, your own data is significantly more understandable to you than what I call OPD - Other People's Data.

Other People's Data is very difficult to use. You have to really massage it to come up with a good results. When you like something on different systems, that means something different on Rotten Tomatoes and on Facebook and on Hunch and on whatever other service you were using. So, you actually get better results with your own data.

RM: The reason I asked that question was I'm interested in what the scale of the ambition is for Hunch. In a recent Wired article, you're quoted as saying that the ultimate goal of Hunch is to "connect every person on the internet to every object on the internet." That is a vast ambition to have...

CF: Oh yes, it is a vast ambition and especially for a start-up that's only sixteen people. But it's important to build very ambitious companies and taking on a good problem, because you want to be able to build a great company.

RM: It seems to be a common theme of everybody I've spoken to in this series. They've all got big ambitions, although their products are also very innovative and forward-looking. So for Hunch, the product changed from the original vision. It started out as a kind of a search product. How do you describe it now, when you're talking to people?

CF: Well, we went from doing search to discoveries. I think that's the crucial difference.

With the prior version of the product, you really had to know what you were looking for. In the new instantiation of the product, we're using the [data from] people who make recommendations in a discovery based product.

Hunch co-founder and Chief Product Office Caterina Fake told Wired in a recent profile that "the ultimate goal of the company is to map every person on the Internet to every object on the Internet, be that a product, a service, or a person."

Why Hunch Exists

The so-called Web 2.0 era of the Web was based on user-generated content and social networking around that. Services like YouTube, MySpace and Flickr (which was co-founded by Caterina Fake) were the success stories of that era.

But now, in 2010, there is too much user-generated content to manually process. What's more, social networking is practically dominated by one company: Facebook. We no longer rely so much on niche sites like Flickr, YouTube, Netflix, Amazon to connect to other people socially. Another aspect to consider is that there's a lot of new data streaming in from sensors, RFID tags and other Internet-connected objects.

The upshot is that we need web services that can help us process all of this data and connect us to the parts that are personally relevant to us.

Opportunities For Startups

The refreshing thing is that these trends are opening up huge opportunities for startups.

GetGlue is another example of a startup aiming to match social data to objects or media. It knows for example that I recently watched Inception and (mostly) liked it. GetGlue can use that piece of data about me, look at my history of other movie likes, connect that to the movie history and preferences that it knows about other people who liked Inception. Ultimately all of that social and 'like' data can be used by GetGlue to recommend other movies to me that I may like to see.

We're early in this era, but both Hunch and GetGlue are busy building up extensive databases about people and what they like (their "taste" data). Not only that, they're slowly perfecting recommendation engines that process this data - ultimately filtering, structuring and personalizing it.

Let us know what other 'beyond social' startups have caught your eye recently.

Log in with your Twitter or Facebook account, answer as few as 20 quick and addictive taste-evaluation questions, and Hunch will turn the front page of the site into a list of highly targeted personal recommendations of movies, books, magazines, computers, meals, vacation destinations and more. It's really impressive.

Hunch is well-known for its "Tell Hunch About Yourself" feature, where thousands of users have answered hundreds of questions each about themselves. All kinds of correlations can be drawn from comparing these responses. People who are comfortable with public drinking fountains, for example, are much more likely to say they'd be willing to risk their lives for a stranger, Hunch has found.

Now Hunch uses that data and more to "personalize the internet." It's a smart change in emphasis.

Where do the recommendations come from? As far as I can tell, the magazine recommendations for example reflect the magazines that people with similar taste as I have (about seemingly random things) and people I'm friends with on Twitter or Facebook have said they like when looking for a magazine.

If you've tried the company's experimental Twitter predictor game, which guesses how you'll answer some very personal questions, based on who you are friends with on Twitter, you've gotten a taste for just how effectively Hunch can look into your mind and soul.

With this relaunch of the home page, the company has cut to the chase and put its most successful features front and center. It's a smart move and one that will make Hunch far more comprehensible and immediately useful.

Co-founder Dixon wrote a blog post six weeks ago on his personal site about how helpful it can be for a startup to pivot. "Ask yourself: if you started over today, would you build the same product?" he wrote.

"If not, consider significant changes to what you are building...You aren't throwing away what you've learned or the good things you've built. You are keeping your strong leg grounded and adjusting your weak leg to move in a new direction."

Taste-gathering startup Hunch is experimenting with a recommendation service that cross references social graph connections on other services with the large set of unusual questions its users have answered. Questions like "do you like facial hair on men? Yes? Well, 48% of our users have said that." The end result is a simple prototype website where you enter a city and your Twitter username and Hunch will show you Foursquare venues it thinks you'll like. Or at least it thinks that people on Hunch who are like your friends on Twitter tend to like those places, on Foursquare. Crazy? Maybe not.

Restaurant recommendations are just the beginning. Hunch knows a lot about a lot of people. The company recently said that the average Hunch user has answered 152 personal questions about themselves. Now that data and our corresponding friend connections are going to be the basis for personalized recommendations. Want to see how well the company thinks it understands you? Check out the recently launched Hunch Twitter predictor game. It's downright eerie.

Hunch co-founder Chris Dixon explained (vaguely) what's going on by email.

We developed the technology to project and propagate our taste data using graph-like connections via public APIs. In this case we propagate our taste profiles to Twitter by projecting the subset of Hunch users connected with twitter onto all Twitter
users. Then we propagate this taste data to Foursquare by projecting the subset of Twitter users checking in on foursquare onto all Foursquare venues. With our collection of taste profiles, in real time we can calculate affinities between any Hunch user, Twitter user, and Foursquare venue. As we project and propagate across all the web's entities, we will enable crazy data mashups. It's going to be cool!

In other words, if Hunch doesn't know about you well enough to make Foursquare recommendations via a Twitter account that's tied to both Foursquare and Hunch, then it will assume you are like those Twitter friends of yours who are on Hunch, and Foursquare.

That's the kind of data-driven value that making all these connections explicit will allow. The future will look like a big algorithm and interface war between companies battling it out to better serve you based on commonly, publicly available user data. Or data you selectively expose in return for recommendations.

This Fall, Zuckerberg got early access to his old friend Adam D'Angelo's new question and answer site Quora and used it to ask: "What startups would be good talent acquisitions for Facebook?"

The winner? Apture.com, the provider of rich multi-media embedded pop-up windows for newspapers and blogs. Founder Tristan Harris is a former Apple engineer who built the first ad server for Wikia, the for-profit arm of Wikipedia, before launching Apture 3 years ago. We've given the product a positive review.

Apture's Harris writes by email: "We we're big fans of Facebook and are super excited about Quora (congrats Adam and Charlie!), but based on the emails that have arrived in my inbox since this article was published I wanted to say that Apture is not for sale. On the contrary we're actually aggressively hiring engineering to join the team and prepping for the release of the next version of Apture. We're totally flattered by the vote of confidence from Quora users, but just wanted to set the record straight."

The next most popular suggestion? Austin, Texas location based social networking service Gowalla. Gowalla is run by CEO Josh Williams, who previously built and sold small business invoicing service Blinksale.

Those sound like good suggestions and both got votes from other Facebook team members on Quora. Remember, this isn't about what technologies should be integrated directly into Facebook - FriendFeed has become little more than an occasional test bed for Facebook feed developments. The question is about scooping up teams of red-hot developers.

Other suggestions offered include Dodgeball co-founder Dennis Crowley's new location based social network Foursquare (it's only a matter of time until Facebook starts doing location check-ins, right?) and social question answering service Hunch, built by Flickr co-founder Caterina Fake and engineering whiz Chris Dixon.

Who do you think would make a good talent acquisition for Facebook? Mark Zuckerberg wants to know.

Hunch is a social Q&A service that, in effect, says, "people who are like you and who have preferences like yours tend to be happiest with the following answer to that question you're asking." The company reports seeing one million unique visitors last month, and in his own blog post about the announcement, Wales calls the intersection of community and algorithm "the future of the web." "This," he writes, "is what we are going to come to call Web 3.0."

In another report about the intersection of food choices and political ideologies, the company says it found the following:

• When it comes to choice of lettuce, everyone likes romaine, but conservatives trend heavily towards iceberg and liberals trend heavily towards arugula.


• For kitchen styles, conservatives vote for the wooden, country look and liberals lean towards sleek, stainless steel.


• Conservatives are more likely to drink sugar soda but less likely to drink wine; liberals are more likely to eat vegetarian options and more frequent portions of fruit.

These questions and answers are ostensibly not the point of Hunch, though. The point is to help users make decisions about things like what blue jeans to buy or what neighborhood to move into. The site has undergone some recent design changes and it's unclear that the main Q&A is as compelling or interesting as the Teach Hunch About You part.

Hunch says it aims to become the "Wikipedia for decision making." The sites are clearly similar: both are user-created and curated collections of knowledge. While that's a laudable goal, I haven't found myself going back to it regularly after our initial review. I'm more of an Aardvark kind of guy when it comes to social Q&A. Perhaps it shouldn't be a surprise then that Hunch says Jimmy Wales and I have a lot in common demographically but very little in common in our ways of thinking.

In addition to decision-type questions, there are lots of fun personal profile questions asked of users - and patterns emerge. Amongst the 79% of respondents who don't mind drinking from public water fountains, for example, Hunch found that they are most likely to be willing to risk their lives for a stranger, always wear bike helmets and say that ethics are more important than success. The 21% of respondents who say they are disgusted by public water fountains are most likely to say they would not risk their lives for a stranger, they rarely or never wear bike helmets, say that if they were one of the seven deadly sins they would be Lust and believe that success is more important than ethics. That's just the beginning of what Hunch says it knows.

The ostensible purpose of Hunch is to help users make decisions. Unlike traditional Question and Answer services that act more like forums for posting on and don't remember the information that's been shared in the past, Hunch is much more systematic. It uses collaborative decision-trees to tell you what decisions other people like you have made in circumstances similar to your own. Hunch tries to avoid keeping people trapped inside options they already agree with by introducing a "wild card" answer to each question.

Should you buy or lease a car? Should you call your parents right now? What kind of blue jeans should you buy? What ancient language should you learn? Should you hire an architect? Those are some of the questions that have already been well developed on Hunch.

The system has a built in feedback loop that helps build out more possible answers to questions, more details to consider in making decisions and a better understanding of people as more people interact with the system.

The site will open to the public at large on Monday - "Early," Caterina Fake says vaguely, she's not running a traditional PR campaign in need of publicity. In addition to users being able to ask and teach about more than 2000 different topics, developers will be able to draw aggregate data out of the system through an Application Programming Interface (API). That sounds like a whole lot of fun.

Here are a few interesting correlations Hunch has discovered about people so far.

People who say they eat fresh fruit daily are most likely to...

• eat breakfast during the week r = 0.229
• spend most of their money on other people r = 0.188
• consider themselves feminists r = 0.181
• not homeschool their children r = 0.172
• use four or more ring tones on their cellphones r = 0.164

People who could not stack their collection of books up higher than their TV are most likely to....

• not be in the middle of reading a book right now r = 0.315
• never buy local produce at a farmers' market r = 0.226
• not consider themselves feminists r = 0.209
• believe in the death penalty r = 0.189
• have never attended a political protest, march or rally r = 0.181

People who were the oldest children in their families are most likely to...

• prefer to have ten to 20 acquaintances than just a few good friends r = 0.278
• have ten or more browser tabs usually open at once r = 0.227
• believe in a "less well known religion" r = 0.180
• believe prostitution should be legal r = 0.159
• currently be in the middle of a book r = 0.138

Middle Children are most likely to...

• not have a cellphone r = 0.259
• have never had their hearts broken r = 0.205
• always wear a bike helmet r = 0.172
• say Ringo Starr was their favorite Beatle r = 0.131

Youngest Children are most likely to...

• prefer having one or two close friends, over a large number of acquaintances r = 0.210
• say that if they were one of the seven deadly sins, they would be Envy r = 0.175
• not currently be in the middle of a book r = 0.137

This is fun stuff, a great example of what can be done with aggregate data analysis. It's fantastic that Hunch is opening up an API right away. An iPhone app for this already compelling service would consume countless hours is someone built it on top of the data Hunch is making available.

Watch for the site to launch publicly on Monday. In the meantime, over the weekend, please remember that the strongest correlation Hunch found among people who say that "Sundays are for working" - is that they also say they have never been in love.