The Big Bad (Filter) Bubble. Really?

Leave a comment

Had the opportunity to listen to Eli Pariser speak at the Shorenstein Center this week, where he gave an updated/small room version of this talk he presented at the 2010 Personal Democracy Forum on the “Filter Bubble.”  (Pariser is a charismatic speaker who does really nifty Powerpoint slides, but if you don’t have 15 minutes to listen to the whole talk, Ethan Zuckerman has written an excellent summary).

Pariser is a political activist who rose to prominence when an online petition that he launched, calling for a nonmilitary response to the 9/11 attacks, attracted half-a-million signatures in less than a month.  As former Executive Director – and now Board President – of, Pariser knows a thing or two about mobilizing people and (perhaps more importantly) raising money over the Internet.

Beware the Bubble

In a nutshell (in case you have not read Zuckerman’s summary either!) Pariser’s concept of the filter bubble is about how each Internet user is now presented with customized content.  If two individuals used the same search term, Google would give different results.  Facebook apparently gives you newsfeeds from friends who share your views.  (It seems to me that I get all my friends in my Facebook newsfeed, which probably means that I don’t have enough friends for the algorithm to kick in!)

Pariser’s concern is that customization is bad for citizenship, because public conversation would end if each of us only sees data that conforms to what or who we are.  He admits that filters are needed, because the information created since the beginning of human history through to 2003 (five Exabyte’s worth, apparently) is now produced every 3 days.  Media professionals have traditionally served as gatekeepers, deciding what the public sees.  While they may have had biases, they were expected to behave ethically.

Pariser shared that engineers at Google had not considered the issue of ethics, because they are convinced that their algorithms are “neutral.”  He argues, however, that code is not neutral but is inherently political.

Code can be neutral

I am with the Google engineers on this one.  For two reasons:

First, filters have always existed.  Our parents decide what to feed us and teach us.  Our friends may decide to withhold bad news.  Our governments have a whole bunch of secrets we will never know about.  (Like who really killed JFK?).  IMHO, one key filter is language.  Simply put, if a message not in Chinese and Spanish, you would have alienated the world’s two largest audiences.  Google’s translation software has in fact done more than most to break down this filter.

[And just to make my point, here is this same line in Chinese and Spanish, courtesy of Google Translate:


el software de traducción de Google, de hecho, hecho más que la mayoría de romper este filtro.]

Second, if we define neutrality as the absence of bias, we can probably agree  (assuming no deliberate shenanigans at Google and Facebook) that code is neutral.  After all, it points some people one way, and others another way.  The code simply tries to help you prioritize – out of the 5 Exabytes – the pieces of data you may wish to see (first).

The real issue

Which brings me to what I think is Pariser’s core point – that balanced views are needed for good citizenship.  Pariser himself is trying to seek out people who have different views from his own.  While that is admirable, one could also argue that it stems from his professional interest as a political activist.  I think, however, that there are many, many more of us who would prefer to remain in our comfort zones.

To use an analogy, a record producer would listen to various genres and different acts.  He/she might personally prefer jazz, but would listen to Britney Spears and Lady Gaga because of the need to keep up with the trends in the business.  The individual jazz listener, however, may have never listened to Lady Gaga (nor wish to).

Ultimately, it is not a bad citizen who does not seek out divergent views; it is a typical citizen.

Pariser suggests that Facebook could add a slider to let us see stuff that’s more (or less) homogenous.  I think that’s fine.  Facebook has lots of stuff I don’t use anyway.


A Wiki on Whacking


Wikipedia is perhaps the grandest and most ambitious example of intellectual collaboration – online or off – to date.  Our assignment this week is to become part of this revolution by editing a wiki.

I have previously made minor edits on Wikipedia anonymously, but as of last week, I am officially a Wikipedian! My username is AyamGoreng!

The wiki I plan to edit is [Caning in Singapore].  While the current article is fairly comprehensive, I think that the summary and section on judicial caning could be improved to provide a clearer, more objective and detailed picture of the topic.

The issue of judicial caning is one of those subjects that border on the “clash of civilizations” debate i.e. Asian values (the good of the general community) vs Western liberalism (the rights of the individual) i.e. the tyranny of the majority (of law-abiding citizens) over the individual (criminal who is being caned).  There are online articles denouncing caning as archaic and barbaric; many consider this practice as a human rights violation and call for its abolition.

[Note: My personal view is that judicial caning as practiced in Singapore is not an issue of human rights.  It is done in the presence of a doctor, whose role is to ensure there are no permanent injuries.  The question is simply whether it serves as an effective means of rehabilitation and deterrence.  But I will save this debate for a later post.]

At the same time, there is clearly some support for judicial caning.  When 18-year-old American Michael Fay was sentenced to jail and caning in Singapore following his conviction on vandalism charges in 1993, there were polls in the US that actually showed a majority of Americans supporting Fay’s punishment.

The Caning in Singapore wiki is largely factual and objective, but it could benefit from references to the Singapore Parliamentary debates, as well as press releases from Singapore’s Home Affairs ministry (which oversees law and order issues).  It is possible that earlier contributors have chosen to dismiss/avoid these official sources as national propaganda, but it should be noted that several other sources cited – like Amnesty International and the UN High Commission on Human Rights – likewise have their own agenda.

The fact is that while the government may have its bias (e.g. to downplay accusations of human rights violations), it cannot afford to publish untruths as its credibility would be at stake.  As such, parliamentary debates and government press releases/speeches provide a rich trove of data that should not be summarily excluded.  I would use the following publicly-accessible sources to add to the wiki:

Separately, while the article generally adheres to Wikipedia’s MOS, I also have some ideas on how readability could be improved.  I think the Wiki could also do with a few pictures.  I found a picture titled “Judicial caning in Singapore” here but am not certain if it is indeed a bona fide picture or if there are copyright issues involved.

I propose to rewrite the summary as follows:

Original (as of 12 Oct 2010):

Caning is a widely used form of legal corporal punishment in Singapore. It can be sub-divided into several contexts – domestic/private, school, reform school, military and judicial.

Of these, judicial caning, for which Singapore is best known, is the most severe. It is reserved for male criminals aged under 50, for at least 30 different offences under the Criminal Procedure Code. Caning is a legal form of punishment for delinquent male members of the military (Singapore Armed Forces — SAF) and this is administered in the SAF Detention Barracks. Caning is also an official punishment in reform schools and as a prison disciplinary measure. In a milder form, caning is used to punish male youths in many Singaporean schools for serious misbehaviour.

A much smaller cane or other implement is also used by some parents as punishment for their children of either sex. This is not outlawed in Singapore.


In the context of Singapore’s law and order, caning usually refers to the strokes of a rotan whip that male criminals between 18 and 50 may receive in addition to prison sentences.  Under the Criminal Procedure Code, the courts may impose caning for at least 30 different offenses; in fact, for several offences – including murder, rape, robbery and drug trafficking – caning is a mandatory part of the punishment.  Younger offenders (between 16 and 18) would receive a milder version of the punishment.

Variations of corporal punishment also used as a punishment for serious misbehavior or delinquency in the prisons and military, as well as a disciplinary measure in schools and other schools.  In private, some parents use a much lighter cane to discipline their young children; this is not illegal in Singapore.

The Cathedral, the Bazaar and the new Web


Read two interesting essays recently.

The Cathedral and the Bazaar by Eric Raymond is a philosophical essay where technical jargon – like the moats surrounding a medieval castle – deter non-nerds from daring to enter.  Which is a huge pity, because its observations apply not only to software development, but are also salient in fields such as business management and public policy, (where, presumably, non-nerds are abundant).

The title of the essay is in itself an instructive metaphor.  “Cathedrals” are built according to a strict blueprint that only an elite few have access to; “Bazaars” on the other hand, grow organically, as each new merchant opens a stall to sell his wares.

Raymond likens traditional software companies to Cathedrals, where access to the source code is restricted to an exclusive group of software developers; the bazaar is analogous to open-source software, where the source code is developed in public view.

The essay basically outlines 19 guidelines for creating good open source software, and this is where some truths which transcend technology serendipitously appears.  Number 11 is my favorite: “The next best thing to having good ideas is recognizing good ideas from your users. Sometimes the latter is better.”  Just replace “users” with customers/stakeholders/public where applicable.

To social science and liberal arts majors, Tim O’Reilly’s What is Web 2.0 provides a seminal reading into how the Internet has developed in the past 15 years, and what it means.  O’Reilly lists Wikipedia, blogging, open source software, peer-to-peer file-sharing (Napster) as the signposts which herald Web 2.0. This was written in 2005 – which is probably the Internet equivalent of the Middle Ages; written today, he would most certainly have mentioned social networking sites like Facebook and Twitter.

Much ado about “Web 2.0”

According to Wikipedia, the term was coined by Darcy DiNucci, a consultant on information architecture, in her 1999 article “Fragmented Future”; the term rose in popularity in 2004, when O’Reilly Media and MediaLive hosted the first Web 2.0 conference, where Tim O’Reilly and John Battelle outlined their definition of the “Web as Platform.” World Wide Web inventor Tim Berners-Lee, however, has apparently called the term a “piece of jargon,” noting that the Web in his vision was all along intended as a collaborative medium.

Notwithstanding, O’Reilly continues to evangelize “Web 2.0”; his essay is available online in Arabic, Chinese, French, German, Italian, Japanese, Korean and Spanish (as well as English).

A Theory of Evolution

There is no question that the web is rapidly evolving, and that it is doing so in a fashion which is probably better described as being bazaar-like.  At the same time, corporate behemoths like Microsoft and (now) Google, which are more like Cathedrals, continue to dominate the scene at any time.

At first glance, the two seem to be in conflict.  For example, would the success of open source software mean the end for Microsoft?

In reality, however, I think that Cathedrals and Bazaars actually complement each other.  Once, at Istanbul’s Grand Bazaar, I noticed that many stall holders would shutter their stalls during prayer times and head for the grand Hagia Sofia which sits down the road.  Likewise, programmers at large software companies may in their spare time, help build Linux in open source collaboration; at the same time, those who dabble in open source development are getting the know-how they need to land that lucrative job at Microsoft.  It may thus well be there the new Web will accommodate both Cathedrals and Bazaars.