1. Select the "Edit Options" Button
2. "Allow" attentiontrust.org to install the extension
3. Click again on the Download button below
4. Select the "Install Now" button in the extension pop-up. Restart your browser.
 


 

Nick Carr on the Death of Web 2.0

Submitted by edbatista on Tue, 2006-07-04 13:13.

Nick Carr points to users' unhappiness with the new Digg-like Netscape (as described by Carl Taylor) and hears the death knell for "Web 2.0 hype." One disgruntled user was quoted in Carl's article:

"I don't want other people voting on what I should read first. I want to see major national news stories and then, if I want to know about entertainment or sports or whatever, I can click a link," the user said.

"I liked having a choice. This new format is awful. What if the New York Times decided to have readers vote on where things should be placed in the paper? What a disaster. If this is how it will be from now on, I'm changing my homepage."

But is this dissatisfaction really an indictment of the underlying concept? Or is it merely an indication that Netscape's implementation is flawed?

Personally, I think it's going to be difficult for a system relying on undifferentiated attention data drawn from the entire user base (whether it's implicit/behavioral or explicit/attributed data) to effectively promote content that's truly relevant, meaningful, and valuable to any single individual.

Attention data (in any form) has the potential to power highly effective discovery and recommendation systems, but their effectiveness is going to be dependent on giving users the option to "tune" them appropriately. On certain (rare) occasions, I'm interested in what the world at large (or at least a particular system's entire user base) is paying attention to, but it's more likely that I'll be interested in recommendations from a more clearly defined subset of that user base: people unknown to me but who share my interests, or people in my social network, or my co-workers, or my friends and family.

I'd also argue that any system relying exclusively on explicit attention data (like votes) rather than implicit data (such as actual reading behavior) is inherently inferior, at best subject to the whims of enthusiastic cranks, at worst susceptible to outright gaming.

Nick's right about the new Netscape's flaws and the limitations of similar user-powered discovery and recommendation systems, but the conclusion he draws is overly broad and premature. Rather than write such systems off entirely on the basis of one failed experiment, let's learn from Netscape's mistakes and build better tools.

tags:

edbatista's blog
There's a key distinction
Submitted by edbatista on Tue, 2006-07-04 18:24.

There's a key distinction within your comment, Paul: "your fellow man" as editor vs. "a group of people he respects" as editor. I'm not particularly interested in the former either, and that's where I agree with Nick and Carl's critique. If I'm forced to choose between my fellow man en masse and the staff of the Times as my editor, I'll go with the pros. But we're going to have a much wider range of options once these systems are in place and we can move beyond crude experiments like the new Netscape--and that's what I think Nick and Carl are failing to recognize.

reply | report as spam
I fully agree, Ben. To live
Submitted by edbatista on Wed, 2006-07-05 11:15.

I fully agree, Ben. To live up to their full potential (and to attract a wider range of users beyond the current base), future versions of these services will have to 1) integrate implicit attention data and 2) allow us to determine whose attention data is relevant for our personal recommendations.

reply | report as spam
The magic will appear when
Submitted by Gabriel Kent on Tue, 2006-07-11 02:27.

The magic will appear when we have social networks comprised of one's agents and not one directly.

I am supposing the infant phase in this editorial shift, which can be more broadly thought of as media mediation (meMe), is exacerbated by the fact that one is performing the editorial by one's own taste -- not by one's own agents by one's own taste -- a large difference in the total possible set-size as well as resolution into such sets.

Simply put, a few power users will always run these scale-free networks...it is in fact the law of scale-free networks ( http://en.wikipedia.org/wiki/Scale-free_networks ). However as we all know, we just are not as good at mining as our agents traversing an object web would be.

So to see the future of this technolonique, one must first see past these baby web2.0 steps to see our agents refining not only what we might like from the raw feeds but from what our likables’ agents have unmasked as well. Couple all that meMe matching/sharing with your people-attention-monitoring/sharing and you will begin to experience meMe magic.

It might be of value for a site like digg to compose simple agents that are randomly created to like certain associated phrases, which traverse very deeply in hopes of digging ‘more’ random links. This might mitigate the typical herd behavior in any close social system.

That said, like now it will still be us.

The idea of choosing between the ‘pros’ and your fellow man is interesting because I presume most of the popular stories were dugg from more popular sources which became popular ultimately through some ‘pro’s format’ – not too mention more and more pros are representing themselves directly in these networks (however, generally without reputation systems their vote is as equal as yours). It seems the current ‘pro’ is just moving further from provider to supplier. Besides, the idea of who is pro in this space is blurring quite readily.

We stand on the shoulders of giants, no reason to suspect our agents will not do the same.

>||;)

reply | report as spam

Post new comment

*
*
The content of this field is kept private and will not be shown publicly.

*

  • Lines and paragraphs break automatically.
More information about formatting options