A six-month window now defines the survival of open-weights models. Nathan Lambert argues that current development cycles must prove their utility before proprietary systems widen the gap further. This pressure forces researchers to optimize efficiency over raw scale. Practitioners must now prioritize rapid deployment to maintain a competitive edge against closed-source giants.