I think that we should still recognize that AutoML will be (and has been) a game-changer for the data science field. That, in addition to the thousands of MOOCs and formal education in data science, is making it increasingly difficult for people to find data science positions. It is becoming increasingly clear that the way for data scientists to distinguish themselves isn't necessarily whether they can build a cool model (which most aspiring data scientists can do) but whether they understand fundamental software engineering principles (which is still lacking among many data scientists).

Finance is for everyone.