All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
Now there's a Web site devoted entirely to making you laugh. And once the technology that powers the site figures out your taste in jokes, watch out.
Jester, the first online automated joke-recommending service, made its bow this week. It's just the place to go to pick up that tasteless joke you'll tell the boss after drinking too much at the office Christmas party.
And you can always return to Jester next week to pick up some more jokes for the clerk at the unemployment office.
Jester relies on collaborative filtering, the same process that's behind Firefly and Amazon.com. Those sites draw on the collected opinions of the current visitor and past users to make educated recommendations. Instead of suggesting a CD or book based on a survey of previous purchases, though, Jester offers a "sense of humor" database.
"We all have friends who send us a steady stream of bad jokes – wouldn't it be nice to have someone who only sends you funny jokes?" asks Jester's project director Ken Goldberg, a UC Berkeley engineering professor best known for his gallery installations that Web surfers control.
Jester begins its routine by giving each user a series of 25 jokes to rate, ranging from mild ethnic zingers to political parodies. Click to the left of the ratings bar if you didn't crack a smile, or to the right if you fell out of your chair. Jester's database matches your sense of humor to that of previous users and presents several jokes your comedy kin found funny. The site then tailors its output accordingly when you enter your username and password on subsequent visits.
"The Web allows large sample sizes where statistical methods work well for recommending films and music, where it’s hard to categorize what you like but easy to give examples," says Goldberg. "We want to see if a similar approach will work for jokes." But Goldberg and his collaborators – students Dhruv Gupta, Mark DiGiovanni, and Hiro Narita – built Jester as much to test the psychology of humor as to experiment with innovative technology.
"Comedians know about the concept of priming: After you hear a funny joke, the next joke will seem funnier," Goldberg says. "If a computer recommends a joke, how will that influence your reaction?"
So far, Jester’s team has found out that humor is not universal.
"Obviously many jokes are culture specific," Goldberg says. "You won't find a joke funny if you don't understand it. The system naturally clusters jokes that appeal to foreigners or to engineers, or to those who are both, for example."
To adequately test their theories, Jester’s creators may need to create an Amazon.com-sized database. The site currently contains only 70 core gags, but users are encouraged to submit their favorite jokes for analysis and possible incorporation into the database.
"Our recommendations will improve as more people try it," Goldberg says. "We are hoping our user base grows quickly, since you can rate new jokes faster than films or books."
