Getting machines to "learn" from what they are doing is incredibly complicated, but it's a problem that, if solved, will be a huge step toward creating genuine artificial intelligence in machines.
This paragraph was buried at the bottom of The New York Times' recent story on machine learning:
The largest class on campus this fall at Stanford was a graduate level machine-learning course covering both statistical and biological approaches, taught by the computer scientist Andrew Ng. More than 760 students enrolled. "That reflects the zeitgeist," said Terry Sejnowski, a computational neuroscientist at the Salk Institute, who pioneered early biologically inspired algorithms. "Everyone knows there is something big happening, and they're trying find out what it is."
Stanford, of course, is the beating academic heart of Silicon Valley. Its graduates - and dropouts - create more tech startups than perhaps any other institution. So it's highly likely that these 760 students, and their colleagues over the next few years, will create the next new wave of hot machine learning companies.
They will not want for employment, that's for sure. Facebook has started a new machine learning unit, as has Box. And there are a small number of startups like BigML devoted entirely to the topic.