Knowledge production is accelerating sharply with generative AI, while the human, institutional, and regulatory systems that turn knowledge into lived reality scale much more slowly. We argue that the field needs implementability scoring, not a pause.