After struggling for decades, researchers have now successfully developed algorithms that can reliably recognize things, and category of things. For example, these algorithms can be “trained” to recognize “automobiles”, and even if you show them a new car, they will “see” what makes a car: headlights, wheels, general shape, etc… and infers that this new object is most likely an automobile. This will become incredibly useful for on-board car computers and could revolutionize safety.
Masayoshi Son argues that the spread of machine-learning, associated with the increase in computing power and the ubiquitous availability of networks will revolutionize our lives and how we interact with technology. Each of us could be connected to about 1000 devices, he says.
He was speaking at an event in Tokyo, where France and Japan have announced the “French Japan Innovation Year”, a collaborative program between both countries to promote each other’s technology and to collaborate on many tech-related projects. French Prime Minister Manuel Valls was speaking at the event, along with several high-profile Japanese politicians.
In conclusion, Masayoshi Son said that he believes that robots would eventually get “smarter” than humans (at least in terms of raw brain-power), but that we should not fear them, and that they would be “friendly”. That, of course, depends on how we build them… but regardless, it is clear that autonomous machines with increasingly higher capabilities are already here, and evolving fast. Alternatively, others are trying to build robots using existing organic brains…