The tiny transistors in microprocessors and other chips fail all the time. Of course, you rarely notice the effects, because they are loaded with redundancy as a contingency plan against it occuring – one little circuit snuffs, there’s a few others to take its place. Naturally, that doesn’t make them any cheaper to make. Self-repairing chips would need less extra circuitry on board, relying instead on clever ways of rerouting their architecture to avoid failures.

Now, I’m all for finding a solution to global warming, and something that works quickly would definitely be good. But I’m not sure that launching sulphur particles into the stratosphere to reflect a greater amount of sunlight away from the planet is a method that makes a lot of sense – the words ‘unforeseen additional consequences’ leap to mind right away.

Much as the idea of performing experiments on disembodied guinea pig eyes is enough to put me off eating for a little while, the data those experiments can produce more than outweighs the gross-out factor. How else could scientists have calculated that the human eye has a connection to the brain that passes data at the same rate as the average Ethernet connection?

Bad news from Russia – the Dnepr rocket being used to launch the thirteen CubeSat projects failed less than two minutes into its flight. Detail on the cause is not available yet, nor on the fate of the payload, but Kosmotras, the State Launch Commitee, is working at recovering the debris and will report as soon as it can.

Why should we be excited about second-generation high-temperature superconductor wires? Because the American Superconductor Corp. has just announced that their latest 2G wires are rated for currents of 140A over lengths of 300 feet – which for non-electricians means that they are capable of being slotted into commercial power grids, an impressive feat for wires a little wider than two human hairs.