For a period on Monday morning, #MediaLabTalk was a trending topic on Twitter — thanks in large part to those who had gathered to hear LinkedIn founder Reid Hoffman speak at MIT, both in person and online. (Trending topics are those that Twitter has algorithmically determined to be the hottest of the moment.)
And that connection between the physical and digital worlds was a major theme of Hoffman’s talk, titled “How to Benefit 100 Million People: Designing and Building Human Ecosystems For Networks and Marketplaces,” where Hoffman, fresh off his company’s initial public offering, spoke about the increase in data that humans are generating on a daily basis.
Hoffman said he is particularly interested in looking for ways to improve his professional networking site by “mashing up” that data — for example, tracking certain skills among individuals’ profiles — to generate useful results, such as identifying industries where one could use those skills, or regions where people with those skills most often live.
“Out of mass amounts of data, you can build interesting products,” Hoffman said. “You actually end up getting really interesting maps in terms of what’s actually changing.”
One of the topics Hoffman has been investigating for years is the way in which people represent themselves, especially online. For a network as large as LinkedIn, with more than 100 million users, it can be difficult to gauge how accurate data is on any given individual’s profile. As a test, Hoffman compared a random sampling of people’s resumes with their LinkedIn profiles, cross-checking skills, and found that in general, people tended to lie less in their profiles, probably due to the fact that those profiles are public.
“When we found people had more than 10 connections, their profile was more accurate than their resume, as a function of the fact that they knew they were in public,” Hoffman said. “People were seeing it.”
Direct to Video: Recorded on July 18, 2011
The video runs 92 minutes.