The Causal Inference Comeback: Why Correlation-Era ML Hit a Wall
For most of the deep-learning era, the answer to “why is this happening in our data?” was “we do not actually care - we care that our model predicts well.” For a wide range of problems, that pragmatism worked. The models did predict well. The business outcomes followed. The causal questions were left to academics and economists. The mood has shifted in 2026. The cases where prediction-without-understanding fails are now visible enough, and expensive enough, that causal inference has moved back from the academic margins to something practitioners need to know about. It is not displacing predictive ML - it is filling in the gap that became unignorable. ...