I think this is a great explanation. The "Ladder of Causation" proposed by Judea Pearl covers similar ground - "Rung 1” reasoning is the purely predictive work

I think this is a great explanation. The "Ladder of Causation" proposed by Judea Pearl covers similar ground - "Rung 1” reasoning is the purely predictive work of ML models, "Rung 2" is the interactive optimization of reinforcement learning, and "Rung 3" is the counterfactual and casual reasoning / DGP construction and work of science. LLMs can parrot Rung 3 understanding from ingested texts but it can't generate it.