Today StumbleUpon announces a complete overhaul of its platform and logo, as it aims to make the site more visually oriented and simpler to use. Now the user profile makes visible all connections, comments, interests, channels, likes, shares, inbox and history. All of a user's thumbs-up are visible through the profile. The new StumbleUpon also features channels, which are essentially sponsored Twitter-like accounts that a user can follow. Plus, the layout looks a lot more like its social network cousin, Pinterest.

The new channels, which allow users to follow content from sites (The Atlantic Wire), celebrities (Tom Hanks, Jim Carrey) and brands (Gilt, Campbell's Kitchen, Audi), mirror more of a Twitter model. For celebrities and brands, this form of advertising is more subtle and, as a result, less intrusive. The new site also makes it easier to keep track of all the pages, photos and videos that a user has liked in the past; the old version only gave users easy access to posts they had shared.

Earlier this year StumbleUpon launched the Explore Box, a tool that's useful for helping Stumblers find topics related to a single topic. If a user is trying to find topics related to "Sleeping Beauty," for example, type in "Sleeping Beauty" and then a list of related interests pop up, such as "Sleeping Beauty Castle," "The Sleeping Beauty," "Sleeping Bag," "Sleeping Bear Dunes National Lakeshore" (OK, the last two were a bit odd, but that's alright). The Explore Box also suggests things it thinks you might like based on the original interest; in this instance, the suggestions are "animation," "France" and "Paris." In the new relaunch, the Explore Box is integrated right into the user experience.

StumbleUpon is a prime example of the read/write web. Why? Because the user literally writes their own "taste graph" by signaling to the service what interests they want to follow. In the e-commerce space, EBay acquired recommendation engine Hunch to do just that - serve up more relevant content to users.

StumbleUpon launched an iPad app back in July. In November, StumbleUpon came to the Kindle Fire.

How do you like the new StumbleUpon? Tell us about it in the comments below.

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.

Our writer Dan Rowinski was curious about how Zite instituted the new feature without submitting an update through the Apple Store. Normally, most updates need to go through Apple as developers resubmit an apps binary for approval. Zite's CTO Mike Klaas responded:

"A portion of our UI is dynamically-generated on our servers (like the article/magazine content), so that gives us a certain amount of flexibility in deploying features. For big updates, we still rely on the app store submission."

Despite being acquired by major media company CNN this past August, Zite is continuing to innovate and maintain their own brand. With the new Sybil update, Zite will be able to do what it does best through multiple profiles.

The new feature comes six months after the startup acquired Discovereads, a book recommendation engine which is something CEO Otis Chandler cited as a sought-after feature among Goodreads users.

"With Goodreads, it's as if you combine your favorite librarian, your best friend, and a database of two million book titles into one person and ask 'what should I read next?'" said Chandler. "We're the Netflix of book recommendations. As members add more reviews and ratings, we keep improving our suggestions for them."

When most people hear "the Netflix of book recommendations" they tend to think of another Internet giant known for its powerful recommendation engine: Amazon. Goodreads says it can provide better book recommendations than Amazon can because it has more data about what people actually like and dislike, as opposed to just purchases, browsing history and ratings.

"For example, we have more than 174,000 ratings of the best-selling 'The Help' while Amazon only has around 4,400," said Chandler.

The site's book recommendations are heavily influenced by each user's book rating history, so people are encouraged to rate 20 books before checking out their suggested reading list. The service is now available in beta to all Goodreads users.

It is an interesting play by CNN. Zite is a powerful news reading iPad app with a lot of excellent functionality. It operates almost like a "Pandora for news" that gives users serendipitous resources based on inputted interest and usage. Yet, outside of being a cool iPad app, Zite is driven by some interesting technology that could be of great use to CNN.

Worio, as its own stand-alone product, did not take off. Yet, Zite pivoted from the search engine to Zite with now-CEO Mark Johnson as an advisor. Johnson will stay on as CEO and run day-to-day operations from the company's office in San Francisco while founder Ali Davar will remain the executive director and Mike Klass the CTO in charge of research and development. Anderson will report to KC Estenson, general manager of CNN Digital.

"When I became CEO of Zite, I was faced with a difficult choice: take venture capital and race to build Zite into a viable business or join forces with a large company and work with a larger pool of resources," Johnson wrote on Zite's blog today. "Despite many attractive paths in both directions, Zite chose to be acquired by CNN. We found a partner who shares our vision and our passion for building an incredible discovery platform."

To Remain Content Agnostic

Johnson also said that Zite will in no way promote CNN, Turner or Time Warner (which owns CNN) content. The platform will remain content agnostic.

"We see a trend developing here, one that is about the future of content discovery, and we believe that increasingly people will come to learn about new types of content based as much upon their personal interests and passions as they do the news of the day," Estenson said in a open letter regarding the acquisition.

In the history of news media, CNN seems to be the perfect candidate to take over a digital startup like Zite. CNN disrupted traditional TV news by launching the first ever 24-hour new network and CNN.com has been at the leading edge of technology and website design for the last several years. CNN now has a wing in its digital division solely dedicated to mobile apps, starting with the iPad and likely coming to other mobile platforms.

Zite had a little controversy upon its launch in March. It was sent cease-and-desist letters from a plethora of media companies that because of the way it stripped ads out of aggregated content. It rectified the situation shortly thereafter and Johnson was brought on board to be the CEO, replacing Davar.

According to leaked information, it appears that the smart technology found in the Siri iPhone application will now be fully baked into the operating system itself. With Nuance's ability to understand natural language queries, iOS 5 will have it all - voice navigation, voice control and voice assistance - allowing users to go beyond simple search and basic actions. When Siri's technology is fully integrated, users will be able to direct their iPhone to actually "do" things, too.

The news of the Assistant feature comes by way of 9to5Mac.com, an Apple-tracking blog, which reported this over the weekend. A source told the outlet that the development of these features was incomplete, however, and may not even be finished by the time the next iPhone ships.

In the screenshots provided by the source, you can see a new toggle button which will let a user switch the Assistance technology off or on. Why would anyone want such a useful feature off? Well, according to 9to5Mac, the assistant also extracts details from the device, including its location, contact information and music metadata, in order to perform its actions. In addition, the speech recognition technology will involve crowd-sourced data, which is sent to Apple in order to improve the service over time. Privacy-minded folks may not want to share their data with Apple, no matter how well it is secured.

In a follow-up post, 9to5Mac also found additional references to Nuance in the iOS 5 SDK (software development kit). The SDK reveals that the technology will be activated by a button, and may display a screen which shows the speech translated into text, similar to what Google's Android operating system has today.

Siri Knows What You Mean. Now Your iPhone Will, Too

Many in the industry expected that Apple was going to make an announcement about its speech recognition and virtual assistance technology at its annual developer conference, WWDC. When the news was a no-show, the theory was that, perhaps deals between Nuance and Apple had not yet been finalized. Given that no official Apple announcement about either Siri or Nuance has been made, everything about when, if or how the technology will arrive in a final product is speculation.

What we do know about the technology itself comes from earlier conversations we had with Norman Winarsky, VP of SRI Ventures, the R&D outfit that made Siri possible. In June, we talked to him about what sort of scenarios deep integration of Siri's core technology could allow for on an iPhone.

Winarsky explained that with Siri, the engine doesn't just understand words, but is smart enough to also understand the "intent" of a sentence. So, for example, it can understand a complicated request like "find me a great hotel in Palo Alto that has a pool and a fitness center," or "find me a good Italian restaurant that serves spaghetti bolognese." It can even be directed to perform various actions, like "remind me to call Sam tomorrow at 3 PM."

An Assistance Layer on Top of Apps?

In the past, Siri had relationships with several services like Fandago, OpenTable, Hotels.com and Menus.com, to provide its answers. Incidentally, many those same services also have mobile applications in iTunes, which begs the question: how far will the Siri integration go? Will iOS 5 not just provide answers, but also launch apps in the process of doing so? That would certainly go a long way in helping users deal with the growing problem of "app overload." With some 400,000 iOS applications now housed in the app store, the trouble is not finding "an app for that," but knowing which one works best for a given need.

With deep Siri integration, we imagine the service could not just direct you to the apps you have installed on your phone, but even suggest apps that could help you in the future. The idea isn't without precedent - after all, Microsoft is doing something similar with "App Connect," an upcoming app recommendation feature found in the "Mango" update for Windows Phone. But Siri wouldn't just be an iteration on Genuis for Apps, it could possibly be the launching point for performing actions within apps.

Although Apple now owns Siri, SRI International is continuing to develop new virtual personal assistance technologies that could also come to your smartphone someday. For example, it's developing technologies that speak back to the user in order to further refine queries, bots that perform reasoning, act on your personal preferences, and much more.

What Apple may choose to do with the Siri technology is still unknown, but one thing is clear: if the Siri integration goes live, the smartphone is about to get much smarter.

Image credits: Assistant: 9to5Mac; Siri

But what's most unique about this app is the way it reformats the text for the small screen. The "floating text" reading experience, which gives the app its name, reflows text originally formatted for the Web for better reading on mobile devices.

This floating text experience has its roots in the technology Scribd introduced last year, when it began offering a way to convert Adobe PDF documents into HTML5. At the time, the company learned a lot about how to transform and rearrange text from one format and one type of interface to another.

In the new native iPhone app called Float, the team leveraged that expertise so that you can now zoom in on text using pinch gestures, and then see that text automatically reflowed for that particular zoom level. You can also scroll left or right in the app, as you would with a book or magazine.

Content Sources

The content in Float comes from over 150 websites, including The Associated Press, Fortune, CNET, HuffingtonPost, Engadget, Entertainment Weekly and People, but it also comes from the information your friends are sharing on social sites like Facebook, Twitter and, of course, Scribd. And you can share content from the app back to those same sites, too.

A special browser bookmarklet lets you save items you find online for later reading on Float.com or within the iPhone app.

A Unique Mobile Reading Experience

Does the idea of yet another "social reading" app leave you dry? This is, after all, a crowded space - filled with apps like Pulse and Flipboard and others like them. But Scribd says its key differentiating factor is its focus on the reading experience, particularly the small screens of smartphones. In addition to the reformatted text, Float offers an Instagram-like "reading styles" option which lets you choose between different formats that work better for different users, like e-ink, high contrast as well as fun ones like "Gutenberg," because, well, just because.

The app itself exists in a unique niche - somewhere in between the social news readers and the "Read It Later's" or Instapaper's of the news consumption world.

In the future, Scribd will look at different business models to aid its publishers - for example, offering them the opportunity to insert ads into the articles in a better format than is typically possible on mobile. It will also attempt to bring in premium publishers offering subscriptions, and is now in talks with newspapers interested in exploring that option.

At launch, Float is available for the iPhone and Web only, but an iPad app and Android version will launch this fall. You can give Float a try for yourself, here.

In Alfred's case, the app uses its smarts to understand the way that people talk about places, and then creates personalized interest graphs that grow and change with each action a user takes and each decision they make.

It seems that the app works, and people like it.

How Did Alfred Get So Smart?

One of the biggest technical challenges the company had to overcome to release Alfred was building its own Web crawlers for the Internet. These bots look for information about places, harvest that data and index it, including recommendations and reviews sites like Yelp, Citysearch and others.

Alfred also learns the language people use when they talk about places, for example, "bad-***" (replace those stars with a word beginning with "a") actually means "good." It's a smart little bot, that Alfred.

When you launch the app for the first time, it walks you through a quick quiz to establish some of your favorites as a starting point. You can exit the quiz at any time, but it's a good way to quickly train the app, we found. It can also learn more about you from your check-in data, but only Facebook Places is supported at present.

The app provides access to restaurant menus, links to the user reviews it found, photos of the business and offers you the ability to call, save, share (email, SMS or Facebook) a place with friends.

As you continue to use Alfred, it takes into account things like the time of day, the day of the week, your location and other signals in order to provide you with the best recommendations for you. You can also thumbs up/thumbs down the recommendations it provides to further train the app.

Future: Deals Integration and More Platforms

In the future, Alfred plans to integrate deals within the application too, but deals would only be one of the signals it takes into account when providing recommendations.  What's most important, is that it finds the right place for you.

Although the app is only available on the iPhone for now, the iPad version will be out soon, and then an Android and Windows Phone version will follow in just a few months. There will also be a Web experience to accompany the app by year-end or early next year which, will offer a more extensive feature set that what you see now on mobile.

We found that Alfred was pretty smart with its recommendations, and it even surprised us by suggesting restaurants we had never heard of and now want to try. You can give Alfred a whirl too, if you like: the app is free on iTunes here.

Flipboard has also added built-in search, LinkedIn integration and has reformatted how the links from Twitter appear. But the company's biggest update is still yet to come.

The newest release of Flipboard is more evolutionary than revolutionary. The most notable change is the improved access to finding and discovery content. In the latest version, tapping on a red ribbon at the top right of the home screen ("More") takes you into a content guide where you can delve into sections dedicated to topics like "Business," "Science & Technology," "Cool Curators," "Art & Photography" and others.

When viewing articles or posts from these sources, a new "Add" button at the top of the screen lets you mark the publication as a "Favorite" for easier access in the future.

Also new is a "Featured" section which Flipboard will use to showcase its partners, the latest addition being The Economist, which will be available at launch. Although many websites can be viewed in Flipboard, the company's partners have the ability to run magazine-style advertisements alongside their articles, formatted in a reading-friendly iPad layout, and track the visits from the application's users.

LinkedIn, Search & More

Another big update this summer release is the new LinkedIn integration. Flipboard is the first company to provide third-party access to LinkedIn Today, a news feed-like look at various industry verticals. 37 major industries are covered by LinkedIn Today, like Food & Beverage, Law and Non-Profit, for example. The end result is something like a trade magazine for following your industry's most important news.

Other updates to Flipboard include the added ability to continue flipping through a magazine or publication after reaching the last page of article, a revamped way to read articles coming from tweets (the article is now given priority while the tweet and its associated actions are moved to the bottom of the page) and the addition of a search feature that returns results for RSS feeds, Twitter and Facebook updates, results from Flickr, Instagram, Google Reader and more. People search is supported as well.

There are several "behind-the-scenes" improvements, too, to make the app run faster and to offer less noticeable user interface improvements in spots.

Coming Soon: Personalized Recommendations

However, what we really wanted to know was when Flipboard would begin to capitalize on the technology it acquired at launch through a startup once known as Ellerdale.

The answer is that Flipboard has already done so. And we'll see more of that technology soon, according to Flipboard co-founder and CEO Mike McCue.

Ellerdale's smart data-parsing algorithms are currently used in the app's search feature and for deriving what's popular on its network, but Ellerdale's technology will be even further baked into the product in the near future.

Will this just be a series of incremental updates, like what we've seen so far? Or does Flipboard have an even bigger launch on the horizon? Both, says McCue. The product will get smarter over time, as new versions are released. But another update, which McCue says is "coming pretty soon" will offer an even better personalized experience involving recommendations.

In the meantime, Flipboard's next big focus is its iPad app, due to arrive later this summer. Versions for Android or other platforms aren't out of the question, but the company hasn't reached any solid conclusions on that front.

The new version of Flipboard is now live in iTunes here and is a free download.

If you're looking for a new app to try while waiting in line for some of those SXSW parties tonight, give PathCrosser a go and see what you think.

The PathCrosser application keeps it simple - it's not a Yelp competitor where users write long reviews of the businesses they've visited, it just pulls in the checkins you already have on hand - those from Facebook Places and Foursquare. Going forward, the app will integrate additional checkin sources, too, PathCrosser's creators Clark Harris and Matthew Simpson told me when I sat down with them this afternoon to see the new app in action.

What's more, PathCrosser plans to work with third-party APIs (application programming interfaces) to pull in other information that would be relevant, like your tips on Foursquare, for example, which could serve to augment the raw checkin data with your personal notes.

The fun part about using this app is that it can be a great conversation starter - bump phones with a friend and, all of sudden, not only do you see each other's travels by way of your checkin history, the app's matchmaking engine tells you whether or not you and your friend have similar tastes and interests.

Plus, for those who don't use location-based checkin services, PathCrosser provides (or rather, it will provide) a Web-based interface where you can pick out the places you've visited or mark them as places you would like to visit. However, that portion of the PathCrosser service has not launched just yet because this startup built "mobile first," as so many today choose to do.

For a first look at PathCrosser in action, check out the video below (sorry for the quality, I need a better camera):

PathCrosser is available here in the iTunes App Store or here in the Android Market.

Note: Video recorded with a Flip Cam, which was provided to me for use during SXSW.

"With their deep algorithmic book recommendation technology, we're going to be able plumb our database of 100 million book ratings from 4.6 million users to find general patterns of the kinds of books people read and to generate high-quality personalized recommendations."

Discovereads calls its product a "taste engine."

"(It) compares how you rate books with how everyone else rates books. When it analyzes the data, it finds ways in which you are similar to someone else and ways in which you are different. It builds a web of relationships between users and books and uses an array of algorithms (sometimes different algorithms work better for different users or books) to learn each person's tastes."

In a note, they were more specific.

"We are using several different algorithms, including a 'restricted singular value decomposition' algorithm, some 'nearest neighbors' algorithms, and an algorithm that extracts the means from the books' and users' ratings, with more secret sauce in the works."

Goodreads expects to have Discovereads integrated between two and three months from now.

Book photo by Brandon Staggs

It's one thing for Amazon or Netflix to recommend movies or other products you might like (that's a huge business), it's another thing for an automated system to tell you where you should go when you walk out the door of your house, what real-world venues you should patronize. That's something a whole lot of companies are going to try to tackle, including Google and Facebook.

Inside every "where should I go?" question there are other big questions, like: What should I eat? What should I buy? What should I do with my leisure time plus expendable income? What should I do with my life? Foursquare would love to be a service that regularly answers those questions for millions of people.

It's a very ambitious goal. It's reminiscent of the question Google has answered, "What Web page should I look at for information about my interest?" As big as the Web is, though - the offline world is bigger, richer and can be more interesting. Win at recommendations offline and your Web application has really done something big.

Real-world place recommendations are higher risk (in terms of expense to the user who makes a purchase, the effort required and maybe social cost) but also much higher reward (in terms of lived experience gained and potential commercial activity) than online shopping or media consumption are. Can Foursquare get recommendations right and capture some of this huge potential?

What Foursquare Brings to the Table

Foursquare first began talking publicly about experiments with recommendation technologies six months ago. Three months ago the company posted a job opening for a data scientist, believed by observers to be someone who would focus on building out the recommendation technology.

Nearly two years after launching, the company now says it is fast approaching 7.5 million users and has recorded 500 million user check-ins. Foursquare doesn't disclose the number of venues it has indexed but one estimate is about 10 million bars, restaurants, parks, stores and other places across the world. If that number is accurate, that would mean there's been about 50 check-ins per venue on average. That sounds like a healthy little data set to analyze for recommendations.

Here's how the company said today that data will be put to use:

Factors Contributing to the New Foursquare Recommendations

1. Your history
2. Your friends' history
3. Your loyalty to your favorite places
4. Your favorite categories
5. Popular places across all users
6. What day of the week it is
7. The time of day a request is made
8. The quality of tips a place has
9. What topical experts have to say.

The idea is pretty simple: tell us what you're looking for and we'll help you find something nearby. The suggestions are based on a little bit of everything - the places you've been, the places your friends have visited, your loyalty to your favorite places, the categories and types of places you gravitate towards, what's popular with other users, the day of the week, places with great tips, the time of day, and so on. We'll even tell you why we think you should visit a certain place (e.g. popular with friends, similar to your favorite spots). You'll find it's helpful for general things like "food", "coffee", "nightlife" (we built in quick access to these searches) and you'll be surprised by what you get when searching for really specific things, like "80s music," "fireplaces," "pancakes," "bratwurst," and "romantic." The more random you get, the more interesting the results get (though be patient with this first release... sometimes we can't find every random thing).

And outside of the "Explore" tab, you'll see some of this thinking starting to surface on the "Me" tab as well. As we started to tinker with our recommendations algorithms, we started to see "expertise" starting to emerge from the data - we're seeing friends that have been to every karaoke place within 10 miles or tried every burger in Los Angeles. The new "Me" tab surfaces some of this, letting you seek guidance from your friends on the categories and places they explore most.

Now, with over half a billion data points, and with every additional check-in and every tip, foursquare gets a little smarter for you, your friends, and the rest of the community. We're already finding this can be just as helpful for finding a brunch spot in your neighborhood as it can be for helping you navigate a new city for the first time.

That sounds great, but the proof will be in the pudding of course. It's great to hear that this complicated problem is being approached with at least nine different factors taken into consideration. <(Above)!--end:nonyt--> Will businesses try to work those factors so that they will be recommended to more potential customers? That day may come.

A Design Challenge

With that many moving parts, there will be engineering challenges for Foursquare for sure - but the resulting experience is what will matter most. Users will be able to know right away if recommendations are for places they have been intrigued by or places they know already and dislike. It's going to take some finesse to really be compelling.

"Machine learning," wrote Joseph Reisinger in a recent blog post titled Why Generic Machine Learning Fails, "is not undifferentiated heavy lifting, it's not commoditizable like EC2, and closer to design than coding. The Netflix prize is a good example: the last 10% reduction in RMSE wasn't due to more powerful generic algorithms, but rather due to some very clever thinking about the structure of the problem; observations like 'people who rate a whole slew of movies at one time tend to be rating movies they saw a long time ago' from BellKor."

Will the Foursquare team be able to look at all the diverse kinds of data it has and thread the needle of the mobile, location-based, game-like, social experience in a way that means users look to it for recommendations of places to go in the offline world? Is this the feature that makes the millions of people who've looked at Foursquare and asked "what's the point?" reconsider their perception of an abscence of value?

"I actually think most people don't want Google to answer their questions," Google's then-CEO Eric Schmidt said last summer. "They want Google to tell them what they should be doing next."

Google is far from the only company that will aim to solve that problem.

Foursquare 3.0 for iPhone and Android should be available for exploring starting late tonight.

My6Sense, the personalized recommendation engine for both iOS and Android, comes to the desktop today with a Chrome extension that brings the my6sense experience to Twitter.com. What exactly is the my6sense experience? my6sense uses what it calls "digital intuition" to determine what parts of your stream, whether Twitter, Facebook, Google Buzz or RSS feed, are most relevant to your interests.

We met up with Louis Gray and my6sense CEO Barak Hachamov last month to take a tour of the new product and discuss the company's plans for personalizing the Web.

"You should be able to go somewhere like CNN.com," said Gray, "and have it sorted according to you."

This is the long and short of my6sense's vision of the Web - a personalized version of the Web for each person, based on how you interact with content. And the company's latest product - the Chrome extension - takes the first leap off the mobile phone and into users' browsers.

"We have a bigger dream of bringing digital intuition everywhere," explained Gray. "The first step is attacking the noise right in your browser."

my6sense's Chrome extension does just that. It adds a my6sense button to your Twitter page, which allows you to view your Twitter stream on the Twitter website as personalized by the my6sense engine. Once the extension is installed and running, it can filter your Twitter stream for the most relevant content from the past six, 12, 24 or 48 hours.

As ReadWriteWeb's Marshall Kirkpatrick noted when he placed my6sense on his "Top 10 RSS and Syndication Technologies of 2010", a key factor for the company is that it is "focusing on monetizing a commercial API" and "that's a good business to be in." Now, the company appears to be playing that angle by showing how its "digital intuition" engine can be applied as another, personalized layer directly in the browser.

As my6sense CEO Barak Hachamov explained, "the amount of information that will compete for our attention in the future will increase by 10x." This, he said, was just the first step.

First Twitter. Next? The Web.

It looks like a good and relatively simple feature. Surprisingly, HuffPo readers responding in comments on the announcement absolutely hate it!

Could objections to using this data be rooted in the ongoing lack of clarity around Facebook's privacy settings? Or did the company just shoot itself in the foot so badly 12 months ago when it made drastic privacy policy changes that people still distrust it today?

The Down Side

Reader comments range from confusion about the feature to distrust of anything associated with Facebook ("I don't have one of those," several people have said) to distrust of recommendations to concern about self-reenforcing political perspectives.

"Great," one frustrated comment read, "put your readers to work for you, is it not enough that we have to deal with your advertisements." Unbelievable!

Another: "If it involves Facebook, you can count me out.

Those are just two comments of many. So far there's not been a single positive one posted! As a technologist, excited about the future of personalized recommendations to assist in discovery, I find this fascinating.

Behind the scenes, though - I believe that the Huffington Post has long been a very data-driven editorial organization. ("See Demi Moore in a bikini!") The company has been experimenting, for example, with attention decay algorithms that move content up or down the front page. Maybe those data-centric practices need to be kept in the background, where readers can't see them and are thus less likely to object.

Problems for Publishers

Other than the unhappy reaction the feature has seen from readers, there are other risks to implementing this kind of technology.

The down side of this for a publisher like Huffington is of course further indebtedness to Facebook. Facebook as identity provider, Facebook as user personalization data bank, it's not customary for publishers to feel comfortable allowing a 3rd party to control all the data about their own readers. Facebook puts conditions on the use of that data too, such as prohibiting caching of the data.

The trade-off is that the ease of use means the publisher gets far more data and far more accurate data. Publishers who collect user data in a vacuum provide little incentive for their readers to input information about themselves that's true - in this case the Huffington Post reader data comes from a context where accuracy is incentivized by discoverability by a person's friends.

For users - it looks like a win. Once you opt-in to connecting your Facebook account and your activity on the Huffington Post, the recommendations appear topical and interesting. That's how it strikes me, but other people sure seem unhappy about it.

First entry into Extension's Labs is something the company calls The Super Awesome Music Blog Finder Thingy ™. Enter your Last.fm username and it will recommend new music blogs that have posted music from artists you've listened to the most over the last 30 days. It's not great, yet, but it could make a pretty great feature once more fully baked.

Unfortunately, the recommendations include too many low-quality spammy blogs, blogs that link to torrents (a little less easy to listen to) and generally need some refinement.

Extension.fm was founded by Dan Kantor, the creator of AOL-acquired Streampad and the feature in Yahoo's Delicious that renders links to MP3 files playable, and invested in this Spring by Spark Capital, Betaworks, Founder Collective (Caterina Fake, Chris Dixon and others) and Dave Morgan (founder of Tacoda and Real Media).

In other words, chances are good that something interesting is going to happen over there. If that includes recommendations based on data acquired from services all around the web and stored in a central repository, that's cool.