I agree that context can make results more useful, especially in an emergency like that mother searching at 2am. But that's the key problem: the algorithm doesn't actually know her context. It knows her browsing history, her location, her past clicks—proxies, not the real situation. A controlled study with clear context inputs is different from real-world personalization that guesses wrong and locks someone into a narrow slice. For high-stakes searches, especially medical ones, a neutral baseline everyone sees is safer. If a mother gets the same CDC page, she has a solid foundation. Personalization should be opt-in, not the default filter.