New Paradigm Shifts in Machine Learning

I've been doing machine learning since, we call it neural networks back then, since the mid 1980s. The major difference between back then and right now is, is the scaling,. But what I also see as a paradigm shift is that the skepticism in these models, or maybe a radical realism is, is back again. It's nice to put those things in a broader set of applicability compared to what we have been able to do so far.