The way towards artificial general intelligence and artifical superintelligence is ahead of us. The road will be something well worth paying attention to, and the way I find myself seeing it form is by looking at where we are with current developments, and piecing it together to project how the future will form itself. That way, the headlines will not be so alluring as to make it seem as though we are around the last bend, and it becomes easy to push back against the fear mongering that makes it seem as though we have already lost the war with the machines. Google released a paper this week, very tellingly titled'One Model to Learn Them All'. The paper details their efforts to build an algorithim that will allow a neural network to train itself to work on upto eight different tasks simultaneously.
Jun-22-2017, 16:15:13 GMT