Even though IBM’s Roadrunner machine got dethroned from its number one spot on the TOP500 list, a research team at the company made some big news this week at the Supercomputing conference, announcing they had “performed the first near real-time cortical simulation of the brain that exceeds the scale of a cat cortex.” In essence, they made a computer act like a brain, or at least a part of one.
The simulation was performed with the “Dawn” Blue Gene/P super at Lawrence Livermore National Lab, using 147,456 CPUs and 144 terabytes of DRAM. Even with all that computational horsepower and memory, the simulation still ran 100 to 1,000 times slower than a real-time cat. Obviously the project has a ways to go before you can replace Tabby with the electronic version at Radio Shack.
Actually, the long-term goal is to create a computer system that encompasses the range of capabilities exhibited by a brain — and I think we’re talking human-type brains now — including sensation, perception, action, interaction and cognition. And this all has to be implemented with the energy-efficiency and footprint (headprint?) of the organic version. The researchers are envisioning an implementation based on low-power “synaptronic” chips using nanotechnology and advanced phase change memory and magnetic tunnel junctions. What they have today is basically a workbench to support that effort.
So unless I’m missing something, the end game here appears to be to create computers that have human levels of intelligence. The business case is pretty obvious: replace people with machines. In the press release, IBM has managed to make it sound mundane:
Businesses will simultaneously need to monitor, prioritize, adapt and make rapid decisions based on ever-growing streams of critical data and information. A cognitive computer could quickly and accurately put together the disparate pieces of this complex puzzle, while taking into account context and previous experience, to help business decision makers come to a logical response.
I’m guessing the first logical response would be to replace any remaining employees with additional cognitive computers. Hmm… I wonder if IBM has thought this through.