Post #1901148

If you want to talk about healthcare access, find a way for physicians to spend their time thinking about diagnosis and treatment, and communicating with patients, not debating a mirage. People "doing their own research" is not new, and there are times when it is truly helpful, but intellectual friction is necessary. The effect of these products is to affirm biases and deceive people into believing they understand, without friction or regard for truth. And once that happens, it is incredibly difficult to talk people down. A person who reads the literature deeply enough to command the language has invested substantial effort, experienced their extrapolations challenged by further reading, and will naturally feel uncertainty in applying their understanding (even when flawed) to specific fact patterns. A patient engaging in good faith (even in dispute) with their physician is revealing their mental model, and the physician can respond accordingly. A patient using an LLM is projecting a mirage, fluency without logical coherence, addictively convincing, but with friction necessary to be able to follow an explanation of why it is incoherent. LLMs have been shown to reduce persistence, so moving the friction to after fluency/mistaken confidence means that even people who would have experienced the friction of learning for themselves will become unreachable after taking the shortcut. Unfounded confidence and quackery are not new, but these products make it more prevalent and raise the cost and expertise necessary to refute. https://www.nytimes.com/2026/04/13/well/ai-chatbots-cancer.html?unlocked_article_code=1.alA.mc0G.UDL2wkgsX05b&smid=url-share