I was very happy when I introduced myself to Jurgen Schmidhuber and his work on Curiosity. He’s the proof that even older geeks can be really cool. Just trust me on that.

As part of his quest to build the “Optimal Scientist” (and then retire), Schmidhuber, an Artificial-Intelligence-slash-computer-scientist extraordinaire,  asks the very human questions of  “What is Curiosity?” and “What does Interesting really means?”. (Have you ever thought about that?).

His response to these questions, although manifested in hardcore mathematical equations and lots of robotics gibberish, is still funny, engaging and comprehensible even for the likes of me.

Schmidhuber basically says that what makes things “interesting”, is really the distance between what we know about something, and the new information revealed to us at the a certain moment, or as we go through a new experience. And this distance, this gap, has to be just right. Too much will result in rejection, too little in all probability in boredom.

So, for example, if I’m a fan of Madonna’s music, and I listen to a new Lady Gaga song, there’s a good chance I’ll find it interesting. (I don’t have to “Like” it, mind you, but still be interested). However, if for some obscure reason I come across an Animal Collective track, chances are I’ll tune it out as fast as I can and put it in my mental “OOCH! NOISE!” section.

How come? (writing on the chalkboard in front of the class):

[Lady Gaga, new] – [Madonna, familiar] = Just right, Interesting

Whereas:

[Lady Gaga, new]  – [Animal Collective, unfamiliar] = Too far, Uninteresting

Ok. So I can’t exactly publish that on Science magazine, but it sort of makes sense, doesn’t it?

For us at Acclair this is fascinating stuff.  Scrumptious almost. Why? Coming next.

. . .

Sweet Appendix:

Lady Gaga:

Animal Collective: