Replies

This link has been provided courtesy of Darwin Central, "The Conspiracy that Cares."

Interesting experiment they have there, excepting that it's not even close to AI. Programming a robot to turn on an LED is a long way from making it feel emotion.

They probably used genetic algorithms for the aritificial neurons. As useful as genetic algorithms are, they are just that: algorithms. Their computational power is not fundamentally superior to conventional algorithms.

Making a machine recognize something it's doing and recognizing something another machine is doing is a simple algorithm. It could be done in hundreds of different ways, literally. The easiest way would be to have the software recognize that the data input is, in fact, coming from the robot's own camera.

Try reading that thread's comments. Many of the posters there are heaping the ridicule upon the project that it richly deserves.

This robot does not do anything other than what it's programmed to do. It recognizes it's own actions versus another robot's actions not because it can, but because it has no choice. It's programmed to do that, so it does. I could write computer programs all day that recognize if the program itself performed an operation if another program performed it. In fact, Windows has been doing this very kind of thing pretty much since it's inception. It's part of the API.