Self-Replicating Robots

ABC News is running a story that self-replicating robots are no longer the stuff of science fiction. Scientists at Cornell University have created small robots that can build copies of themselves. Here is a movie demonstrating the self-replication process. And the paper that will be published in Thursdays issue of Nature.

Comments

That's not self replication

[catbutt]

Hell they might as well consider the raw material to be "robots that are powered off", and then have the bots push the power button on the "raw material" to create a new robot.

Lame.

Re:That's not self replication

[datafr0g]

Why is this not self replication?
Who said that replication must involve the original robot to create the robot parts? And even if it did, it would still have to create these "spare parts" from smaller parts anyway...
The robot is replicating itself from it's own basic building blocks from what I can see.

Re:That's not self replication

[Andy Gardner]

I agree how is this different to robots on an assembly line assembling a car, change 'car' to 'copies of the robot performing the assembly' and you have a /. story. The only reason it hasn't been done before is theres no point?

Re:That's not self replication

[FleaPlus]

Why is this not self replication?

Indeed. Last I checked, humans and other animals couldn't self-replicate either, but needed to have raw materials preprocessed by things like plants first.

Re:That's not self replication

[andrebasso]

The thing is - humans gather and harvest those raw materials on their own and are completely responsible for the use of those raw materials around them. Actually, it is very close to self replciation. This video shows no harvesting or cultivating of those discreet building blocks in any way. They merely pop into the frame. Very, very far from self replication in any way.

Not exactly "gray goo"

[localroger]

So they can assemble spare parts into copies of themselves. Where do they get the spare parts? Oh right.

Re:Not exactly "gray goo"

[Clod9]

That was my thought exactly. The interesting advances will come when someone creates a process that a computer can control that takes some simple raw material (like plastic resin) to produce new parts, with the design of the new parts under the control of the machine itself.

I envision a factory in which molds are created using rapid prototyping technology, purely from machine-produced 3D parts specifications. Initially, these designs could be hand-created by humans, but automated modifications could certainly be done and with complex enough design software, parts could be created and assembled in a fully automated way.

Think of a drive mechanism that uses four wheels, but testing shows that it needs more wheels to support the weight; the rear axle could be lengthened and a wheel added on each side, and the heavy part of the load could be shifted rearward. This kind of design improvement isn't simple to codify, but then the software used for routing paths on a PC board has more complex rules than these.

This is coming, it's only a matter of time. I give it 20 years before it's applied commercially.

Re:Not replication

[MankyD]

It may not be creating itself from what most would consider "raw" materials, but from its own world view it is. It has a few fundamental building blocks from which it can create more advanced structurues - copies of itself in this case.

I have a better design

[exp(pi*sqrt(163))]

Rather than have robots made out of prefabricated cubes, why not prefabricate the entire robot. Then when a robot wants to reproduce it just has to say "make it so" and lo! and behold! there's another prefabricated robot sitting there. I don't see that this is any less reproduction than this example. Of course, if you use the log probability measure mentioned in the paper it doesn't score too well but that could be fixed by giving each robot an on/off switch that another robot can press.

I'm sure I've seen more bogus papers than usual go by recently.

Re:Edward F. Moore's 1959 self-reproducers

[Anonymous Coward]

I thought it was a strained tour-de-force then, and I think these "self-replicating robots" are just a fancier example of the same thing.

We are just fancier examples of the same thing.

Re: What about the SG-1 team?

[ArielMT]

Ah, yes. The unstoppable replicators, which could only be defeated by, as the asgard Thor put it, human stupidity. No self-replicating form of artificial intelligence can stand up to natural stupidity.