Scientists have built autonomous nanobots powered only by chemical energy that can "sense" their environment and repair broken circuits.
Adam Rifkin stashed this in Nanotech!
Tiny autonomous robots don’t need computer programs to repair circuits too small for the human eye to see:
Not all autonomous robots need artificial intelligence to power them. At the molecular level, nanobots can do pretty impressive things without lines of code dictating their moves. They do our bidding because the physical laws of their environment force them to do so.
By exploiting such quirks of nature, scientists have now built nanobots that can repair broken circuits that are too small for a human eye to see. Such tiny repairs could help modern electronics have a longer shelf life, but these proof-of-concept, autonomous nanobots have bigger potential. They could one day soon be used for self-healing materials and delivering drugs inside the human body.
To build them, Joseph Wang of the University of California at San Diego and Anna Balazs of the University of Pittsburgh took inspiration from nature. When you cut yourself, the platelets in your blood sense the wound and start aggregating to start the healing process. They wanted to create tiny robots that could do something similar.
So they started with Janus particles made of gold and platinum. These spherical nanobots (or “nanomotors” as the researchers call them) are thousands of times smaller than a pinhead and have two surfaces with distinct properties. This choice was critical to power the nanobots to act as Wang and Balazs wanted them to.
Top Reddit comment:
Okay, I have to set the record straight on this. First let me be clear that this is still incredibly exciting stuff, but the article (as well as the authors' choice of words) is causing people to draw conclusions that are a little more fantastic than the research shows. The particles are not autonomous robots that sense their environment and steer to repair the circuit. They simply shoot off in random directions. The ultrathin crack is energetically favourable for the particles to attach to, so particles that hit the crack stay, while they do not attach to the surface in other locations. This is being called "sensing", however it is not a long range interaction at all, the particles do not know about the crack until they run into it. Furthermore, the effect would not work if the crack was larger, say even a micron. This technology heals ultrathin scratches and surface defects in the metal, and it can't fix your circuit unless the problem is specifically a nanocrack.
In conclusion, nanotechnology is awesome and these researchers developed extremely clever technology, however you don't need to worry about the nanorobot apocalypse quite yet :)