I’ve been using different AI models a lot, and I’ve noticed that newer versions of ChatGPT seem more careful and restricted than before. Even normal or creative requests sometimes get refused or answered in a very safe way. At the same time, I see more people talking about using other models because they feel more flexible and actually helpful for everyday use. Do you think OpenAI is striking the right balance between safety and usefulness, or do you feel they’re leaning too far into restric...