Slashdot

Could this algorithm be applied to porn?

These are funny once, Mike, not funny always.

"I don't understand, Man."

It'd be really interesting to use a site like this to try and determine how people's sense of humor (and response to jokes, which isn't the same thing) clusters.

It's hard in this case though because the jokes are so old and tend to fit closely to five or six templates. Because this means they have very little impact, I tried rating them based on how funny I thought they would be if they were new to me and expressed a bit more concisely, which I found a complicated exercise.

If this thing could be loaded with new jokes, or at least varied jokes, it'd be very interesting to observe the results. for example, would we find that people who like gender-stereotype jokes also like lawyer jokes? Would we find that people who like engineering jokes also like pun-based jokes?

Alas, without a system for users to submit their own jokes I don't think there's enough data in the system to get useful results out of it.

P.S. Shakespeare walks into a pub. And the barman says, "Sorry, you're bard."

P.P.S. So this bear walks into a bar, and the bear says "I'd like a......... beer, please." And the barman says "What's with the big pause?"

P.P.P.S. So this woman walks into a bar, and asks for a double entendre. So the barman gives her one.

I rated 8 jokes and it gave me this recommendation:

Could not connect to user database: Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (11)

I don't get it.

I must say, after ten jokes, it got me spot on. Longer, story-like jokes with ironic or twist ends! Me likes!

In my case it seemed to hit on all the things i hated first and then after about ten jokes in, just started riffing. I wonder how long till i exhaust the database...

It must have real problems with political jokes. Half the people think they're real funny, while the other half don't. And that half changes by joke!

They are trying to develop the funniest joke in the world.

Before the slashdotting I made it through the 'training' and then through about 20 jokes, it wasn't too bad.

And don't worry, the jokes get longer it's only the initial ones that are the one or two liners.

But before it went it actually did seem to be homing in on my sense of humor.

A castaway crawls up on a far-way beach only to be greeted by a hundred angry, armed savages. The chief approaches him.

Castaway:

Oh, God, I'm screwed!

Suddenly, the clouds split asunder and our castaway hears the voice of God:

No, my son. Thou art not screwed yet. Pick ye up the rock before you, and bring it unto the head of the chief.

Inspired, the castaway lifts the heavy rock before him and smashes it into the skull of the chief just as he walks up. The chief falls down dead.

Again the voice of God booms:

Now you're screwed.

Underlying it is a Berkeley-patented "collaborative filtering algorithm" dubbed Eigentaste , now on Version 5.0. The more people who use the system and rate jokes, the more data Berkeley researchers have to advance their understanding of recommendation systems, like those used by Amazon.com and other Web sites."

Yes, of course I want to give Berkeley researchers my time for free, so they can add that to the public subsidies that pay for their research, so they can patent the technology and charge me to use it later.

I'm the punchline!

If this story had been posted tomorrow, April 1, I would have dismissed it out of hand as some sort of hoax. April Fool's Day is the worst day of the year at /. One of my New Years' resolutions was to not visit /. at all tomorrow. It's just not worth it and, unlike some of those Jester jokes, the stories aren't even that funny.