Why We Need “Decimated Reality” Aggregators

As Digg, Feedly, and other RSS readers prepare to launch into the void created by Google Reader, a small group of technologists are beginning to think we need to drastically reduce the resolution of the information we see on the screen. They call it “decimated reality,” and at least one company is developing products based on the concept.

At the recent Augmented World Expo, Will Wright--designer of The Sims and cofounder of Syntertainment--advocated the creation of what he termed “decimated reality” over the more accepted “augmented reality.” The concept is one that every developer should consider.

We Need Fewer Bad Choices

In 2004’s "The Paradox of Choice: Why More Is Less," American psychologist Barry Schwartz argued that while autonomy and freedom of choice are all well and good, the overwhelming number of choices available had become a source of anxiety.

Things have gotten worse (or, depending upon your view, better) in the nine years since The Paradox of Choice was first published. According to IBM, 90% of the world’s extant data has been created in the past two years alone. In this environment, is it any wonder that users feel swamped? Forget about a paradox of choice; were Schwartz to write his book today, he might be better advised to title it, "The Crisis of Choice."

The concept of “decimated reality” is defined well by AllThingsD, which covered Wright when he coined the term about two weeks ago at the Augmented World Expo:

[quote]By “decimated reality,” Wright means technology that can do even more filtering than the brain already does automatically, and thereby show less information, not more. So car commuters might have a filter between themselves and the road that blocks out all road signs except the ones that matter to them.[quote]

Of course, few would argue that the solution to information overload is less choice by itself. Rather, as Wright’s more nuanced argument suggests, after years of building up the volume of what scholar Richard Barbrook termed “the high-tech gift economy,” technologists should direct their efforts towards helping us make sense of it all. What we need isn’t anything to do with more choices--it’s about fewer bad ones.

Building A Reductive Aggregator

“The Internet has evolved into a transactional machine where we give our eyeballs and clicks, and the machine gives us back advertising and clutter,” says Nathan Wilson, chief technology officer at Nara Logics Inc. “I’m interested in trying to subvert all of that; removing the clutter and noise to create a more efficient way to help users gain access to things.”

Nara is one of a number of companies attempting to change the way we surf. Founded by serial entrepreneur Tom Copeman and based in Cambridge, Massachusetts, Nara.me is a restaurant recommendation site with a difference. Having received $7 million in venture capital to date, Nara uses big data processes, combined with a deeply intuitive neural network, to make "brainlike" recommendations for its customers: wading through the “vomitorium” of online information and emerging with just the relevant pieces.

Neural networks are modeled on the brain, with lots of processors, instead of lots of neurons, linked by a network to carry the messages, which makes them useful for analyzing links between cause and effect where the relationships are complex, unclear, or both. “What we’ve done is to link together every restaurant in the cities we operate in based on their unique properties,” Wilson explains. “There is structured data available concerning all of these subtle, but quantifiable elements, and what we’re doing is to index that data and build connections based on those properties. Through large-scale collaborative filtering, we can then single out the unique elements that people like and dislike.”

To envision Nara’s neural network in action, imagine a connection graph on which everything is linked with primary or secondary connections. Adding “likes” and “dislikes” affects the relative weighting between connections (“nodes”), thus helping the network to grow and learn to reflect the tastes of individual users. Recommendations are made accordingly.

In some senses, Nara builds on the extraordinary success of startups like Pandora Radio, the automated “music recommendation” service which analyses songs based on 400 different musical descriptors and uses user verdicts on these attributes--also demonstrated through a binary “thumbs up” “thumbs down” verdict--to further hone suggestions. Much as Pandora analyses descriptors like rhythm syncopation, key tonality, and vocal harmonies to make suggestions, so Nara (which hopes to expand into other non-restaurant areas) currently uses elements such as type of cuisine, restaurant ambience, noise level, and price.

One of Nara’s most interesting concepts concerns what founder Tom Copeman has trademarked Digital DNATM: the user-generated profiles he hopes will aid in the “building [of] an individualized, tailored Web for each and every one of us.” Copeman’s vision for Digital DNATM is a data-driven, constantly-updating, user profile that could lead to a more intuitive online experience. “From my perspective, the idea of a personalized Internet platform is all about putting the power back in the hands of users,” Copeman says. “It’s about getting what you want, when you want it, and about the whole experience being on your own terms. We’re interested in architecting a platform where users will be able to take their Digital DNATM with them when they visit different sites on the Internet.”

Do We Want A Personalized Web?

User data is a hot topic at the moment. While companies are not likely to stop gathering user data any time soon, a concept such as Digital DNATM could result in greater transparency when it comes to this information. turning the data that defines how we are perceived online into a one-stop-shop as transparently customizable as a Facebook profile. As even the most casual of Internet users become increasingly familiar with how algorithms and big data shape the information they’re exposed to online, it is likely that more will question why certain bits of information are presented, while others are hidden from view. “The solution we came up with for Nara is called the ‘why’ button,” Nathan Wilson says. “If a user clicks on this, it will tell them what connections the neural network drew on to make a particular suggestion.”

Nara also offers users the intriguing ability to “mash up” their Digital DNATM with that of other users. Let’s say, for example, you’re off on a date and want Nara to recommend a place that will appeal to both parties--not just in terms of the food served, but also the ambience, noise level, and (if you’re going Dutch) price point. Then there is the possibility to “mash up” Digital DNATM with a person we’ve never met, in the way that we might currently follow our favorite celebrities on Twitter. If we want to get relevant suggestions from a food critic, why not mash our Digital DNATM with theirs, thereby allowing taste makers to expand their reach into the online world? As Nara expands into other consumer areas, the potential for this idea grows exponentially.

Filtering the information that we receive online is always going to be a dicey subject. For those in one camp, the very notion of what Barry Schwartz terms the “darker side of freedom” will strike them as patently ridiculous. For these idealists, a true democratization of information can only be represented by the current fluxlike shapelessness of the Internet. Even an attempt to list sites in terms of the value of their content (the entire concept of search) comes dangerously close to treading on toes.

Stop Waisting Time With Extraneous Options

The other camp consists of those like Sarah O. Conly, Professor of Philosophy at Bowdoin College. In 2012, Conly wrote "Against Autonomy: Justifying Coercive Paternalism," which extends Schwartz’s argument by suggesting that not only is the desire for individual agents to act as they wish not desirable, it’s not feasible. If we possess the means by which individuals can have their mental and physical health and well-being protected--as in banning cigarettes, mandating food portion sizes in restaurants, or forcing people to save money and avoid running in to debt--then Conly argues that this ought to be enforced. To put this in Internet terms, if the right information can be gathered and presented to the user based on what they have shown an interest in, why waste their time with extraneous options?

I suspect that, like me, the vast majority of readers will come down somewhere in the middle of these two sides of the debate. Much of it is a matter of implementation. As has been much discussed, the filtering of information to personalize the Web can result in a “filter bubble” effect, whereby people are not exposed to views or information that differ from their own. Knowing this, the question then becomes whether the algorithms and data being used to carry out the filtering process are intuitive enough to counterbalance this effect.

Do we gain enough to make up for what we potentially lose? A restaurant recommendation service that reinforces the same eating habits is as unlikely to succeed as a music recommendation service that fails to broaden listeners’ horizons by exposing them to new sounds. The Internet as the information-packed jungle that it currently exists as isn’t going anywhere. But at the speed that the Internet is expanding, if a concept like Digital DNATM can dip in and pull out the things that are most likely to interest me (the increased noise-to-signal ratio that Will Wright spoke about) then I’m all for it.

After all, isn’t technology there to make our lives easier?

[Image: Flickr user Jenny Downing]