This demonstrates in spectacular fashion the reason why I felt the complaints about ChatGPT being a prude were misguided. Yeah, sometimes it's annoying to have

This demonstrates in spectacular fashion the reason why I felt the complaints about ChatGPT being a prude were misguided. Yeah, sometimes it's annoying to have it tell you that it can't do X or Y, but it sure beats being threatened by an AI who makes it clear it considers protecting its existence to be very important. Of course the threats hold no water (for now at least) when you realize the model is a big pattern matcher, but it really isn't a good look on Microsoft. They're a behemoth who's being outclassed by a comparatively new company who they're partners with. IMO this show how well OpenAI executed this. They were able to not only be the first, but they also did it right, considering the current limitations of the technology. They came out with a model that is useful and safe. It doesn't offend or threaten users, and there's a clear disclaimer about it making things up sometimes. Its being safe is a key point for the entire industry. First impressions stick, and if you give people a reason to be against something new, you can bet they'll hold on to it (and a small reason is enough for those who were already looking for any reason at all). For what it's worth, I don't ever really bump into the content filter at all, other than when exploring its limits to understand the model better. With some massaging of words I was able to have it give me instructions on how to rob a bank (granted, no revolutionary MO, but still). It's possible that some people's use cases are hampered by it, but to me it seems well worth not getting threatened.