Monday, May 10, 2010

Harvard Admitted Students Weekend Review, Part 2

As I mentioned in the first part of my review of the Harvard Law School Admitted Students Program, I'm a big fan of Jonathan Zittrain, the founder of the Berkman Center.  Thus, take this discussion of his presentation with a grain of salt.  Even my acquaintances at HLS knew that I was geeking out throughout the whole thing.

JZ started out with a playful jab at the Yale University Press graphic designers who were creating the image for his latest book, "The Future of the Internet--And How to Stop It."  Their first two attempts were emphasizing the stop aspect--with a hand and then stop sign--so he decided to crowdsource the project and ended up with the incredibly compelling image to the right.

In order to introduce his further musings, JZ talked a little bit about the Socratic method, how he employed it while teaching, and then jumped right into some background information we'd need in order to be able to intelligently discuss some of the issues he'd present.  The first of these concepts was part of his original story, crowdsourcing, which he used as springboard to start his presentation on "Minds for Sale."
That link to wikipedia.org is probably the most self-reflexive link I've ever made.  JZ lists Wikipedia as one of the positive examples of crowdsourcing, opining that in a perfect world, an academic institution like the Berkman Center would have started it.  One of the things that is so exciting about Harvard Law School is that it interprets its mission to not only study and prepare its students for the legal field, it also wants to play a part in shaping it.

JZ moves on to describe a pyramid of possible tasks that could be "crowdsourced," organized by increasing skill and concomitant decrease in number of people able to perform the task.  At the top of the pyramid he placed events like the recent DARPA challenge to find 10 weather balloons across the United States.  The winning team from MIT utilized online social networking in order to pool the resources and combined localized knowledge of people across the country.  This type of time-sensitive fact-finding has obvious implications for a variety of Homeland Security-type purposes.


At the bottom of this task pyramid, JZ places such simple tasks as typing appropriate keywords that are relevant to a given image.  Ever heard of Google?  While many corporations have clearly identified the benefits of crowdsourcing, JZ and I agree that there is a disturbing lack of knowledge from the general public about how these types of automated, seemingly robotic, computer functions are actually performed by humans.  And for free!


For instance, for the image above I offered, "polar bears, cute animals, snow love, bear hug, polar bear hug" and, of course, "anthropomorphizing love."  The more complicated your tags that are matched, the more points you receive.  The only thing the people who "play" the Google image labeller game receive are points.  Which can be redeemed for nothing but geek points.

While those types of tasks have interesting applications--and are helping Google's bottom line, a noteworthy observation in and of itself--JZ is more concerned with a notch in the ladder just above free.  I'll get into more detail about that next week.