●researcherWorth watching because it frames a concrete limit on what scaling and automation can contribute to scientific methodology versus conceptual breakthroughs.
●policyRelevant to arguments about AI's role in accelerating science — the paper provides a formal basis for distinguishing capability growth from genuine discovery.