Deep learning, the final stage of automation and the end of work (again)?

This paper reviews developments in the field of AI to focus on deep learning, a field and set of applications that has recently experienced an explosive take-off after years of slow progress: deep learning as cybernetics in the 1940s-1960s based on biological theories of learning; deep learning as connectionism in the 1980s-1990s; and deep learning since 2006, not necessarily neutrally inspired but rather a type of machine learning that allows computer systems to improve with experience and data. This paper explores some of the dimensions of the recent phase of deep learning in terms of the final stage of automation, exploring its implication for education and work.