In a recent lecture at Imperial College London, Stuart Parkin — research fellow at IBM’s Almaden Research centre, and the guy who perfected the spin-valves that make modern computer hard disks work — discussed some of the challenges faced in trying to build artificial an brain. He quoted some amazing things about the differences in scale and complexity between computers and brains.
According to Parkin in terms of raw computational power we now have supercomputers that are capable of performing an equivalent number of computations per second to that of the brain of a rat.
But a rat achieves this in the space of less than half a cubic centimetre. And it consumes about 50 milliwatts. Whereas IBM’s BlueGene supercomputer is massive. It takes up 600 square feet. And requires 400 kilowatts of power.
Obviously, we’re going about it the wrong way. But what is the right way. How do living systems create the magnificent complexity and shear power of the brain?
We asked neuroscientist Richard Wingate.