An article in today's Times explains that, in the next year, "learning" computers will become commercially available. The cutting-edge software in these computers is modeled on the human nervous system. By imitating the human brain's ability to develop new neural pathways in response to external stimuli, these computers will effectively "learn" from their mistakes, thus eliminating (or at least greatly reducing) the need for reprogramming.
The breakthrough moment? "Google researchers were able to get a machine-learning algorithm, known as a neural network, to perform an identification task without supervision. The network scanned a database of 10 million images, and in doing so trained itself to recognize cats."
From that point on, computer scientists posed ever-more-challenging identification problems to these learning computers, which demonstrated exponential improvement. The computer quickly identified
images of George Clooney:
And Gwyneth Paltrow:
And finally, when tasked with procuring an image of the Grand Canyon, the computer responded in a matter of nanoseconds:
Truly, the future is upon us!