Harvard is getting computers to think like humans

Modern computers are powerful data-processing machines, but they still need a tremendous amount of information to accomplish certain feats of recognition. Researchers at Harvard University are working hard to rectify this situation. Computers have grown from bulky room-size contraptions that could do simple calculations, to supercomputers capable of simulating extremely complex tectonic movements, involving 1 quadrillion floating-point operations (FLOPS) each second. To illustrate the matter even better, here's a great article that explains a 1 trillion-fold increase in computing performance through history. Then again, even though our brains can't compete with the sheer speed with which computers churn out results of even the most complex equations, when it comes to things we take for granted, like creativity, abstract thinking or pattern matching, machines fall way behind.