Victus Spiritus


How do you define and capture relevance across the web?

18 Mar 2010

The breadcrumbs were eaten

In the previous incarnation of (in development, formerly you could see what your friends (or a lists) main topics discussed were in a moment by typing in the URL /users/twitterusername. Now with the expansion of potential information streams being "the web", there's no guarantee or meaning in /users/victusfate being me (unless I claim that identity on our site). We've had to rethink the way we handle flowing information, and how best to connect that to a user identity.

The biggest challenge for shifting the social tools at Victus Media away from a fading legion of twitter-apps, and becoming a world wide web application has been maintaining relevance. It's clear that the url breadcrumb doesn't have the same instant meaning as it once did to the community of twitter users. But identity methods, and the sources of web information we can process has expanding immensely. Where once we could potentially cater to 5 million active users of social media, we now can engage over 1 billion & growing possible web users.

The trick to making the transition is staying focused on relevance, and expand on a simple way of creating order from the chaos of information overload.

We're not like any other social reader/aggregator

At Victus Media we believe in self curated input channels. But being able to direct one's own info stream doesn't mean you can't borrow the ubiquitous attention of super human filters. Generous folks like Louis Gray, Robert Scoble, Marshall Kirkpatrick, and Mahendra Palsule create numerous public web streams that you parse with our filters. We extract entities, or the most important people, place, ideas from as many feeds as we can handle. Simply add your favorite feeds as sources you follow. Our plan is to make it so that users can identify

There will be a feeds overview page to add, delete and organize feeds. We'll work out the details with helpful user feedback.

Mobile and location are additional sources of info we can use to aid in the neverending search for user relevance. The more orthogonal data we can get, the better our suggestion algorithms can eventually get. Our longer term vision leads us down a path that makes a highly functional web agent that can negotiate great deals of data and highlight relevant serendipitous sources browsers may normally miss. The same agent can be used to make each web browsing adventure adapt to our preferences. As our interests change the web agent will shift it's criteria and learn from our recent behavior.