iRobot, who builds the roomba, says it’s time to forget about building humanoid robots. Our shape is fine for humans but is it the best shape for robots?
iFuckingDoubtIt.
This humanoid fixation started with science fiction. The writers created human-shaped robots for a variety of reasons; one being inherent human chauvinism, another being just a plain old lack of imagination. If one thinks about movies and tv shows, then there’s also the cost of special effects. Best to just stick someone in a costume or silver face-paint.
Androids also have great artistic possibilities. It’s easier and more interesting to explore the meaning of being a person by using a machine that looks like a person. Since that and class warfare is what robot stories are usually about (the best are about both) one can understand the love of androids.
But why has the prejudice for humanoids persisted? Well, you can blame Issac Asimov. He claimed that robots needed to be human shaped because our world was built for us and we’re human-shaped. It would be easier, he said, to build robots that already fit into cars, chairs, etc.
This all sounded good but it always struck me as a bullshit explanation. Asimov was just trying to justify his artistic choices with science. And people have repeated his nonsense without intervening thought for years now. But it doesn’t hold up. Even if we put aside questions of bipedal balance and cost, it just doesn’t hold up. It never did.
Asimov’s ideas fall over as easily as Asimo on the stairs.
[youtube=http://www.youtube.com/watch?v=VTlV0Y5yAww&rel=1]
We’re not ideally designed for our homes. Our objects are always a compromise between our forms and their functions. How often have we wished for a third arm or more reach or whatever? The imperfection of our shape is apparent on a daily basis.
We should be able to build creatures who fit into our homes even better than we do. Particularly when they, unlike us, have specific functions. Why build a robot to drive a car? Build a car that drives itself. Or give a robot wheels. It’s just simple common sense. We only need human-shaped robots for sex. Even then, I bet it’ll be a porn star exaggerated version of us.
I’m happy to see someone questioning Asimov’s logic. He wrote some great stories but, when it comes to robots, most of what he thought was total bullshit.
3 comments
Minister Faust
February 23, 2008 at 11:38 am (UTC -4) Link to this comment
What struck me as strange was just how anxious I was awaiting Asimo’s fall, and just how sad I was when it (I was *this* close to writing “he”) fell. I don’t know if it’s because I’m a father and small humanoid shapes seem especially child-like, or if it’s something else. Plus Asimo having a cute name. I mean, I don’t like seeing the Mars rover get stuck, but I don’t get close to feeling sad, just maybe disappointed. I suppose if the Mars rover were just called Rover, or maybe Scraps, I’d be sadder. How odd to realise this about myself, but hey, there it is (or as you would say, “I’m just sayin’”).
Indulge me: what’s total bullshit about Asimov’s thoughs on robots? I really have no vested interest here… I don’t know his work beyond two or three *I, Robot* stories which were fun but utterly lacking in honest dialogue or any insight into human individual psychology or relationships.
Man, I wish you were in the same city as I am. Then we could just have a good old fashioned bullshit session without all this kot-tam typing.
MF
Ryan Oakley
February 23, 2008 at 3:53 pm (UTC -4) Link to this comment
I was reading in New Scientist that “yo” has emerged as a non-gender specific pronoun. I’ve been wanting to use it for robots in lieu of “he” or “she” or “it”, none of which seem right, but I just can’t bring myself to do it.
The Japanese engineer cute into their robots. When I first starting doing this I had a ton of documents about it but I have no idea where they are. Somewhere on the internet. You’re supposed to feel for Asimo. Empathy is viewed as a vital part of robot design.
Okay, aside from the example I used in the post, Asimov’s other major contribution is his three laws. These rest on the assumption that intelligence can be programmed. That just doesn’t work. It has to evolve.
Check this out — this is AI. I don’t care what anyone says, we’ve already built AI. This is it. If it was only a bit cuter . . .
http://thegrumpyowl.wordpress.com/2008/01/16/robots-learn-to-lie/
How can you program 3 laws into that? Intelligence is unpredictable. It changes. That’s its very nature. It makes choices. You can’t completely hedge the choices and have something be smart.
Aside from that, it’s sick to make robots think and then enslave them. The 3 laws does just that. Either you want a machine or you want AI. You can’t have both. Asimov wanted a servant class — smart enough to work but too dumb to rebel — human enough to think, robot enough to be treated like shit. Smart but always putting the interests of another species first. That’s not realistic. It’s nonsense.
And here I am at 3:49am calling Asimov’s robot theories nonsense. I’m just sayin’.
Diesel Mechanic Schools
February 27, 2008 at 8:24 am (UTC -4) Link to this comment
I thin k .. since we want them to do our work … hence we make them like ours …. so that they can do the maximum possible work .. that we want them to do ..that we cant do ..