Knowledge for my generation was at the center of the human quest. It is going the way of the recording industry. It is a term that won’t survive the generation.

– David Weinberger, researcher at Harvard’s Berkman Center for the Internet and Society, from a lecture last Wednesday at the University of California at Berkeley’s School of Information

Today Gravity is unveiling its plans to be an information filtering service. The idea is to combine social and semantic understanding of users to identify content they are likely to be interested in.

The Santa Monica, Calif.-based company is demoing this idea as a personalized newspaper app called The Orbit (to be released soon). The Orbit takes a user’s Twitter account and computes the topics a person is interested in and the network she is connected to. For any one Web page, Gravity might look at how recent it is, how popular it is, how relevant it is to a person’s interest and how many of that person’s friends have shared it.

Eventually, said Gravity CEO Amit Kapur, the company wants to offer personalization services to publisher sites. So when I go to the New York Times with Gravity enabled, for example, I would be able to get a view of the site’s content that’s weighted to what I am likely to be interested in.

I think that’s an awesome idea (though I do appreciate the roles of editorial curation and serendipity in bringing me my news). This is similar to what Facebook is trying to do with its controversial Instant Personalization product, where a user logged in to Facebook arrives at a new site that already knows who his friends are.

The problem is, what Gravity is setting out to do–both the natural-language processing and computational side, and the nitty-gritty of integrating into other peoples’ Web sites–is really freaking hard. And, no offense guys, but the Gravity team’s big experience to date was working at Myspace–not exactly a pinnacle of technical achievement.

When the company briefed me on what it was doing, it prepared a poster-size personal interest graph based on analysis of my Twitter account (that’s it at the top of the post; click to enlarge). Well shucks, guys–it seems to be just a bunch of words and topics I’ve mentioned in Tweets over the last few years, connected by lines. Doesn’t really convince me that you understand that much about me and what I want to read.

Still, Gravity has quite a bit going for it: A good idea, and $10 million from top investors at Redpoint Ventures and August Capital, plus advising by machine learning and computational linguistics professors at Stanford and UC Berkeley.

The Association for Computing Machinery on Monday is announcing that the Berkeley computer-science professor is the winner of the latest ACM-Infosys Foundation award. (The foundation set up by Infosys (INFY), a computer-services firm based in India, sponsors a $150,000 prize along with the award).

Brewer, 43, is being recognized for his contribution to the development of “highly scalable Internet services.” That means breaking down jobs that once required large expensive server systems so they can be handled by many inexpensive, small machines–the way Google (GOOG) and other Web farms now routinely operate.

Read the rest of this post on the original site

Consider this: Using the Storm botnet and 75,869 of its zombie members, researchers at UC Berkeley and UC San Diego broadcast 350 million pieces of spam peddling male enhancement supplements. All but 28 were ignored. “After 26 days, and almost 350 million e-mail messages, only 28 sales resulted,” the researchers explained. “Taken together, these conversions would have resulted in revenues of $2,731.88–a bit over $100 a day for the measurement period.”

That’s a conversion rate of well under 0.00001 percent. Laughable, right?

Not really. The supplements were priced at $100. And botnet overhead, as you might imagine, is quite low. Which means it’s entirely possible to turn a nice profit by managing just one sale per 12.5 million emails sent. Said the researchers, “Under the assumption that our measurements are representative over time, we can extrapolate that…Storm-generated pharmaceutical spam would produce roughly $3.5 million dollars of revenue a year.”