Mike wrote about this yesterday, but I've gotten a few emails today so wanted to add some comments.
Researchers at Cornell have developed some cubes with electromagnets that can move to connect themselves together. They have called this self-replicating robotics -- which it is, in a sense. But the components are far more complex than the "replication" operation.
In terms of ideas, it's very significant that self-replicating robots are newsworthy and mainstream.
In terms of technology, this has almost nothing to do with molecular manufacturing. In MM, the fabrication operations will be a lot more information-rich than the feedstock, enabling general-purpose manufacturing.
It's a stretch, but a useful one, for the researchers to mention nanotechnology. But that does not warrant this National Geographic story talking about gray goo and buckyball toxicity -- two topics that are unrelated to each other, to current molecular manufacturing designs, and to this robot research. If I weren't about to hop on a plane to Europe in a few hours, I'd write them to complain about this confusing reportage.
It should be mentioned that electromagnetic fastening, the method chosen by this group, does not scale to very small devices. So in terms of nanoscale manufacturing, these robots are a distant proof of concept only.
But... nice work. A small step in the right direction: general-purpose manufacturing via repeated digital operations on small components.
Chris
You're right, the National Geographic story is pretty bad. They ought to know better -- and you'd think they could at least spell Ralph Merkle's name right!
Posted by: Mike Treder, CRN | May 12, 2005 at 08:38 AM