Interesting thought process Mikhail! I think not having domain knowledge or just the ability to apply data science through logic is a huge gap in the industry right now. Most people assume the role of a data science is to run R, SQL, Phython etc. scripts and build accurate ML models. The reality is that unless and until one can marry the results with business outcomes and actions, data science is just a piece of information, A few year ago, I briefy worked with a "genius" data scientist who told me he built a logistic regression model with 98% accuracy. It took me 5 mins to look at his variables and ask few questions to realize that he had used an independent variable that was a pseudo for the dependent variable. It could have been picked up in correlation analysis, but we seldom run correlation b/w dependent and independent variables. However, having clear understanding of data and business acumen/ logic is a must. It's not "Or', it is "and". So, to your point, I feel data scientist will not be extinct, they will evolve to understand that data science is not just coding or data manipulation or running complex ML algorithms. It's about understanding the business objectives and using data to get the right insights to arrive at the best possible actions. Apologies for such a long comment, I am just passionate about this topic ;)