Brace yourselves, humans: we’re one step close to the Singularity, as researchers in Spain have built and programmed a robot that is learning to recognize itself. Meanwhile, the so-called geniuses at Harvard have built KILL-BOTS… I’m sorry, wait, it’s actually “kilobot.” As in: they’re so tiny, that thousands (kilos) of them can be programmed to do various tasks. Still, they might want to work on that name.

Robotics company The Corpora built an adorable, trash can size robot named Qbo entirely on the open source Linux operating system. One of their main goals was to give Qbo effective object recognition skills. In a video posted recently to The Corpora’s blog, Qbo accurately identifies a (Linux mascot) penguin, but is then presented with himself.

As many have pointed out, Qbo technically doesn’t have “self-awareness”, certainly not on the level of humans. He simply identified an image, and learned to associate a word (“me/myself”) with that image. If Qbo were to meet another Qbo, he might also think it was “me.” Still, the research seems to be a step towards artificial intelligence.

The Kilobots, on the other hand, let thousands of tiny robots work like one. According to the press release from Harvard’s School of Engineering and Applied Sciences, the bots are each the size of a quarter with three “toothpick” like legs, and cost only about $15 each. They can be programmed by the thousands. But once programmed, they can work autonomously, coordinating the thousands of individual movements themselves.

Says Dean Michael Mitzenmacher: “Tiny robots are really cool!” Hasn’t he ever heard of the Gray Goo theory?

So we’re making robots self-aware, and letting them coordinate their efforts by the thousands. What could possibly go wrong?! What do you think?

[Photo courtesy Michael Rubenstein]

Bookmark to: