Kate Compton

Kate Compton

Tell us about your research...

My research work is on something that I’m defining as casual creators, which is a genre of creativity software that nobody has identified yet but a lot of examples of these exist. The opposite of this is sort of professional creativity software. Things like Photoshop and Garage Band where you have to be able to have total control over the thing that you’re making because either you’re making it for a boss or you’re making it for some sort of grand vision.

It turns out there’s this entire other genre of software that’s for casual creators, people who are using this thing autotelically, which means for its own purpose. So you’re playing a pottery wheel app, not because you want pottery, but because you want something to do that’s not Angry Birds. So there are little apps where you’ve got maybe 15 seconds to get people to start thinking really creatively and then you have to have them enjoy the experience. So I’m drawing a big circle around all these apps and there are literally thousands of them on any given app store, and figuring out what it means to make creativity apps for casual use. How do you make an AI into a kindergarten art teacher instead of a soldier or all the other evil things that AI can be? Basically, how to make an AI Bob Ross.

What brought you to UC Santa Cruz?

I used to work at Electronic Arts on a game called Spore, which is this giant sort of Manhattan Project of, what if we could procedurally generate a whole world and all the creatures in that world and all the plants in the world and all the galaxies as far as the eye can see? This brought together a whole bunch of people doing really fascinating work who would have never worked for EA before.

I was a particle effect artist on Spore so I was doing planetary generations making particles move around a sphere and drilling out mountains and laying down rivers behind themselves. I used to call myself a geological choreographer because you’re kind of making this choreographic dance of particles. Then when there were no more projects to do along those lines they had me doing just kind of blood poofs and laser poofs and lots of little shooty weapons effects and I could kind of see my future at EA being the next 20 odd years of making little weapons poofs.

At the same time I was coming to UC Santa Cruz to Michael Mateas’s lab because Michael had been my advisor from 2004-2006 when I got my Master’s at Georgia Tech and then he left as I was writing my thesis and came here to start building up this department. He’d have me come down here and judge student competitions and I was doing guest lectures at a couple of the little conferences that were local and I could see all these really exciting things being done at UCSC and it made it pretty easy for me when I decided I wasn’t growing anymore at EA to come down here, join Michael’s lab and refuse to leave for the past eight years.

Tell us what project(s) you’ve been working on...

In my Interactive Narrative class there was one assignment that was called “story grammars” and this is kind of the idea of making a little mad libs for the computer to run. For example, for something like a Scooby Doo mystery, the van breaks down in some way, a mystery is revealed, some sinister things happen and then they unhood the person and it turns out that was the real estate investor.

You could kind of make yourself little mad libs for what that would be, like different ways for the van to break down, different spooky things to happen. I was learning JavaScript at the time so I made myself a little JavaScript language that would do that and then I released it Open Source, that was a language called Tracery and that got picked up by somebody in the U.K. who then made a platform to host Twitter bots on it. That became wildly successful so now I’m the proud grandmother of roughly 8,000 Twitter bots that we know of that are running on a particular thing, unless Twitter has banned all of them in the last week which they’re kind of working on.

It’s open source so nobody is required to tell me anything about what they’re making with it. So we hope none of the sinister bots but we’re not sure. But a lot of really fantastic different art experiments, it’s DJ’ed a London music event, which is I think the most DJing that any of our lab projects have done. What do you like about the Baskin School of Engineering? All the professors here are so supportive. It’s wonderful that we have a faculty here where nothing is off limits. Believe it or not, there are many computer science faculties out there that, if you said you wanted to do fabric, they would look at you funny and say that’s not really computer science. No one has ever said to me that’s not computer science or even in our department, that’s not a game.

What’s your favorite bot?

My favorite bot, because it’s proof you don’t have to be technologically advanced in your bots, it’s called InfiniteScream and all it does it post different variations on “AH.” So the grammar for this one, the little recipe is not complex, this is something that anybody can make in about five minutes. But because it’s such a reliable bot...it screams three times an hour, I believe, the person who made it keeps upping the number of screams as news happens. It’s been declared to be the “Bot of 2016,” “Bot of 2017,” and “Bot of 2018.” If you tag the bot in conversation it will scream back at you and so I will frequently encounter this bot.

I don’t follow it just because it screams so very much, but I’ll often encounter it when I’m reading a news article on Twitter that somebody has posted about something that’s happened and they’ll say, “Infinite Scream, what’s your take on this?” And it’s kind of the ultimate character actor that you can bring into your little improv piece on twitter.

Computer Science
Degree Program: 
Computer Science and Engineering
Place of Birth: 
Atlanta, Georgia
Undergraduate Institution: 
Pomona College
Graduate Institution: 
George Institute of Technology
Michael Mateas