He hints that Shardanand's start-up may be looking at software that places different levels of "confidence" in articles, based on the author's reputation. It's an unproven idea, but that doesn't mean it's impossible. It could be a user-voting system (like eBay's ratings); a method based on the most-linked-to people (like Google News, except for individuals); or an approach that uses collaborative filtering, sending a user a "liberal" or "conservative" version of a story based on the articles she's chosen before (like TiVo and Firefly).

A long and controversial article at SF Weekly complains about Craigslist's consumption of the newspaper classified ad business, and then goes on to provide a number of details about Newmark's attempt to create something an aweful lot like MemeStreams.

sfweekly.com | Craig$list.com