The necessity of causal models is a paradigm shift

Causal diagrams are a form of optimized information compression. Causal diagrams crystalize knowledge, make it more transmissible, more accessible, and reduce evaporation of information.

The necessity for causal models is a paradigm shift that collides with the prevailing AI/ML meme of digital culture. The “causal revolution,” like all real revolutions, will be bumpy and full of friction. I think the resistance to Pearl (see “bashing statistics”, or “this book is a failure”) reflects, and is proportional to, our ‘automagical’ fantasy. And our emotional attachment to cognitive ease. It is my impression that the greater part of the resistance to ‘Why’ may come from those beguiled by the promise of AI/ML relieving us of complexity and the onus of cognitive effort. Those invested with the status quo, who identify with the prevalent ‘data-centric intelligence’ or with conventional statistical practice will also be offended. This is natural behavioral economics: bounded rationality and ‘satisficing’; and is to be expected.

Our becoming better scientists (health scientists, data scientists, computer scientists, social scientists, etc.) will not progress without ‘a push’ (extrinsic information). ‘Why’ is a cause, of progress.