By Jared Newman | Friday, August 6, 2010 at 1:16 pm
Just because IBM’s Deep Blue beat Garry Kasparov in chess 13 years ago doesn’t mean computers are always better gamers than humans. The scientists at University of Washington in Seattle learned this first-hand, when a program they created for protein folding got its best results from real people.
Science journal Nature has the fascinating story of Foldit, a free downloadable game in which players fold protein for the benefit of science. It was originally a computer program that would run in the background, much like the [email protected] project, but as users watched the program meticulously fold amino acids into more durable three-dimensional shapes, they complained about how slow it was. Foldit was created as a way for people to speed up the process, and it encouraged gamers with high score tables and the ability to form strategy groups.
This week, the University of Washington announced that the best Foldit players indeed excelled over machines. That’s because humans take risks and have long-term vision, neither of which are strong suits for computers. By determining the most efficient way of folding amino acids into more durable three-dimensional shapes, gamers are helping scientists target the inefficient protein arrangements that cause allergies and neurodegenerative diseases.
Obviously, not every instance of crowdsourced science is conducive to gaming, and as Nature points out, there’s a fine line between volunteer work and exploitation, but I like the idea that gaming’s competitive and puzzle-solving qualities can be put to good work. If only Farmville was so productive.