The dominant paradigm in AI research and development envisions the goal as automation of tasks and entire jobs. But another, comparatively marginalized vision has existed since the beginning of AI — that of augmenting workers and amplifying human potential. In this talk, it is argued that the automation agenda is struggling…