Thanks for reading John. And how much data science work should I have done before I’m allowed to express an opinion? You misunderstand, I never said AutoML will replace data scientists, I said the ability to code cannot be the thing that defines a data scientist. I disagree on supervised being easy vs Unsupervised learning, supervised is not easy. Each has its own challenges. Ease depends on the quality of the data or the problem you’re trying to solve, not the tool.