For one thing, I ran into an interest in Common Lisp on the part of the younger set. There are several manifestations in use. One called clojure has been used for both front-end and back-end work. What is suggest to me is that we can start to talk a workbench approach.
Aside: There is a lot to discuss, but Lisp has been involved with AI from the beginning. Here is some information related to why it is so good: What-did-Alan-Kay-mean-by-Lisp-is-the-greatest-single-programming-language-ever-designed.
If one listens to some AI/DS/ML/DL (that is, Artificial intelligence, data science, machine learning, deep learning) practitioners, one sense can an out-of-control situation. Gobs of time, money, space get gobbled running against suspect data, one might say. As, this information is wide and not very deep. So, who is deep? In other words, where is the science?
Too, there is a general malaise with regard to not understanding. That is, the computer comes up with something, and the humans cannot provide anything that allows comprehension to look at it and make a judgment. No, the reaction is to think the the computer is out of our league.
That might be true in several senses, but it is not in terms of truth (hence, we need to engineer this).
Remarks: Modified: 11/29/2017