Lowell Goss - Comments

I like the idea and the 3R names is a sweet piece of marketing, but the UI seems a bit lacking. Wallpaper on the sites is distracting and you only get maybe 3 stories above the fold. It's hard on the eyes and from a glance it doesn't seem like there's much "there" there. - AJ Kohn

13 more comments

sweet - Anthony

AJ, you would be correct, so far. They've got some work to do. - Louis Gray

But they are on the right track -- massive statistical data mining to find the most important signals in the noise. - Sean McBride

@Sean: Agreed, if they can really crack signal/noise and do it across many verticals ... watch out. - AJ Kohn

What's most interesting about LOUD3R, despite the fact they're not polished, is their array of topics. TechMeme obviously appeals to an audience that SNEAK3R readers wouldn't see, for example. And Ballhype has already shown me that a sports-focused site like this is good, but LOUD3R has broken them into individual sports. We'll see how this one does. - Louis Gray

why is this "massive?" - Jason Goldberg

@Jason: Because there are 550 sites, or will be. 550 is a massive number of sites. :-) - Louis Gray

Interesting, would like to know more about the rules engine. - jcunwired

Jason: I should have said they are trending in the direction of large-scale statistical data mining -- the more sites the better (and especially the more dimensions of sites the better). "Massive" more aptly applies to Google and datasets of billions or trillions or items. - Sean McBride

@Louis: agree on reason for massive. If they even get this half right, they have a powerful network of identifiable sites across numerous verticals, some well served, other not. (oh, found one they don't have!) - AJ Kohn

Great comments. I will try to answer a couple of the questions. - Lowell Goss

The rules engine for determining "Most Interesting" has about 15 different key parameters. These breakdown into three areas; Quality, Community and Buzz. The weightings are varied from site to site giving a different "editorial voice" to each topic. - Lowell Goss

Lowell: is there a technical paper on the rules engine available on the net? Or is that information proprietary? Do you have any thoughts on the rules for determining "interestingness" on Friendfeed? - Sean McBride

We do not currently have a plan to publish the details of the "interestingness" calculation. In regards to the application of the concept to Friendfeed...I think that the key parameters are similar. How does the software evaluate how interesting an individual story is? Is the story text only or does it include media? Are users interacting? On Friendfeed this might be a measure of the quantity and velocity of commenting or sharing. - Lowell Goss

Lowell Goss › Comments