The current version available at https://chat.openai.com/chat seems neutered to the point of being both less useful and less interesting than just searching google.
If they don't do that people goad the model into recommending we eat glass or similar ideas.
Surely we shouldn’t throw the baby out with the bathwater, no? Trying to control this is surely futile — how long until a company like Stability AI releases an open source, uncensored version of GPT?