Causality as Robust Prediction

I like thinking about causal estimation as a type of prediction (e.g., here). Here's a very nice slide deck from Peter Buhlmann at ETH Zurich detailing his group's recent and ongoing work in that tradition.