The FDA finalized guidance for apps related to drugs, setting an updated compliance and expectations framework for digital therapeutics and drug-linked software. For developers, the decision narrows uncertainty around what must be regulated and how evidence is expected to be generated. In parallel, a separate debate focuses on evidence standards for generative AI-enabled medical devices, with healthcare lobbying groups and tech companies weighing in on how much validation the FDA should require. The tension centers on balancing safety and clinical utility with the operational realities of rapidly evolving models. Finally, a set of usability and implementation studies highlights persistent friction: prototypes can show generally good usability in healthy volunteers while workflow failures remain unresolved—underscoring that adoption depends on connected system reliability, not just model performance.