In a move that mirrors the oldest tension in the distribution of powerful tools, OpenAI has simultaneously widened the gate and tightened the rules — offering its latest language model, GPT-5.6 Luna, freely to all comers while imposing stricter guardrails on what the system will say and to whom. The company, serving roughly a billion weekly users, is attempting something genuinely difficult: building a single experience that serves the curious student, the professional analyst, and the vulnerable teenager without compromising any of them. It is a wager that access and responsibility can be eng