This is a much-needed perspective on the new generation of tools in language modeling, object recognition, robotics, and others. The time and effort spent pitting algorithms against human intelligence is truly mind-boggling, when algorithms have been complementing us in so many tasks for decades. The new generation of tools simply offers more opportunities.
In data science, for example, humans excel at conceptual modeling of causal problems because they are creative and imaginative, and algorithms excel at complementing effect identification by collecting, structuring, computing, and optimizing high-dimensional, high-volume data in nonlinear, nonparametric space. Maybe we just need to get over the obsession with benchmarks that pit machine against human and create tests of complementarity.