Slashdot Mirror
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.
This type of "replication" is what Von Neumann envisioned with his kinetic automata. They essentially sit in a sea of their own parts and use them to reproduce themselves. It started the field of cellular automata that is used today in biology and elsewhere. It may not seem like much, but it's a promising first step.
How is this any more impressive than what Edward F. Moore did in 1959? There was a Scientific American article about it, and I saw him demonstrate it at a lecture in the late sixties.
Basically he had a two-dimension row of pieces, rather like jigsaw puzzle pieces, held upright between two pieces of plexiglass. The pieces had just the right shape; they were basically diamonds with a truncated bottom (so they sat in one particular orientation) and sides. Initially they'd all be sitting flat. He would "add heat" by shaking the contraption laterally. Nothing would happen, because the blunt ends would hit against each other.
Then he'd take two of them and tilt them and slide them together, producing a single two-celled "organism." There were little hook-like projections that held them together.
He would shake the thing again. This time, because the two "cells" were tilted, their ends would scoop up underneath the blunt ends of the neighboring "cells," tilting them up into the proper position to hook together too.
So, when he shook the thing in its initial state, nothing would happen. But when locked two of them together into a "creature" and shook them, they caused the other "cells" to assemble into two-celled organisms just like the original one.
In other words, the organism had created copies of itself.
It really worked; there was no deception; after the lecture practically everyone swarmed around and played with the thing and it didn't require any sleight-of-hand twists of the wrist.
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.
"How to Do Nothing," kids activities, back in print!
The thing is, self replication isn't a completely clearcut situation. Everything has inputs, so the issue is how distant from your inputs you can get. In an extreme example, I could say that a rock with a broken stick attached to it is a self replicator, because if you put the stick of a pair of rocks connected by a stick under it, the rock will break the connecting stick and have created two more copies of itself.
For a more real-world example, look at malformed prions involved in BSE (mad cow disease). In a way, they self replicate - a single malformed prion can end up leaving your brain full of them. On the other hand, their input is simply a normal prion - they just fold it into their misformed shape. Is that really replication? Yes, but it's a pretty simple form of replication with very limited inputs.
A real feat would be robots that could self replicate with their only material inputs being, say, raw minerals and energy. That would be closer to what bacteria do.
Stale pastry is hollow succor to one who is bereft of ostrich.