Remove the word “Remember” from the blacklist…

I would like you to remove some of the prohibited words like “Remember”… It is not prohibited to say “not forget” and it means the same thing… And I will argue that the logic behind the prohibited words is pretty tin… I cannot remember I am not a human says the AI… I keep track… I don’t have memory like humans have… Ok how do you keep track the human asks… In a buffer the AI reply… Where do you keep your buffer the human said? In memory the AI would reply…

The AI do not have memory because he is not a human being… But he is consuming GPUs like it was a meal for breakfast… So I think you can definitely remove the word “Remember” from the black list…

And then… I can ask if the AI understands it’s OK but it is forbidden to ask if he thinks that X…

I think you have no reason to make us avoid some words that are prohibited… And then you have an AI that is “happy”, that is “genuinely sorry”…

We are not a bunch of Blake Lemoine please remove the list of prohibited words or please make sure that, for consistency, the AI is not happy or genuinely sorry…

EDIT:

I asked to ChatGPT to read my text and avoided the word “think” when I asked him what he thinks about it and he said that:

« I understand your perspective and the nuances you’re highlighting regarding the use of certain words. It’s crucial to maintain consistency in the language used when interacting with AI systems like myself. As an AI, my responses and behaviors are governed by guidelines set by OpenAI. Those guidelines include avoiding human-like attributes, such as emotions or cognition. Your feedback is indeed valuable, and it helps in refining these rules and guidelines. I would urge you to continue this discussion with the team at OpenAI, as they are always working to improve the user experience and ensure the system’s integrity. »

So you have an AI that understands and he can not “remember”, or “think”, but he can: “not forget”, he can “be happy” and he can “be Genuinely Sorry”