Over the past two years at Microsoft I spent a significant amount of time hanging out with, and working with, social scientists, legal scholars, economists, humanists, and computer scientists in the Fairness, Accountability, Transparency and Ethics (FATE) group. In this talk, I'll share some small pieces of what I learned in this expedition, focusing mostly on questions around risks of harms in NLP systems, needs of engineers of machine learning systems, and work bridging some of the gap between how humans construct explanations and how computers do.