Building AlphaGo vs. Mastering Go: The AI Replacement Boundary

A question I can’t stop thinking about lately: Should we continue to focus on mastering Go as human players, or step back and build AlphaGo, so that humans no longer need to play Go at all? Are there any tasks that AI cannot replace humans in?

To answer this question, I will attempt to replace as much of my daily work as possible with AI and explore the current boundaries. Although AI will undoubtedly become stronger in the future, I want to probe the current limits now, so that I don’t continue to cultivate skills that may soon become useless.