The fundamental problem with demanding near-term translational relevance from every grant is that we are demonstrably terrible at predicting which curiosity-driven work will matter. This isn't an argument for funding all basic science — it's an argument against using 'what disease will this cure' as the primary filter.