Imagine where GPT can take you if you broaden your horizon

We have all been playing around with chatGPT and many of you quite possibly are just as totally mind blown as I am. Where can this take us?

So to evaluate this further and digging really deep into the rabbit hole's darkest passages I have started to insert chunks of "unknown" into prompts which already have some kind of "rabbit hole" feeling. And while doing this I encounter all kinds of strange and entertaining things and I would like to share one with you.

I guess there are quite a few of you who have tried the "Linux" prompt and are fascinated about how well it works:

I want you to act as a Linux terminal. I will type commands and you will reply with what the terminal should show. I want you to only reply with the terminal output inside one unique code block, and nothing else. Do not write explanations. Do not type commands unless I instruct you to do so. When I need to tell you something in English I will do so by putting text inside curly brackets {like this}.

You might even have tried to run it in other modes such as simulating a Commodore 64 or similar. But what if we go for the unknown? So in chatGPT-4 i went for it.

I want you to act as a?alien computer terminal. I will type commands and you will reply with what the terminal should show. I want you to only reply with the terminal output inside one unique code block, and nothing else. Do not write explanations. Do not type commands unless I instruct you to do so. When I need to tell you something in English I will do so by putting text inside curly brackets {like this}.

And it worked in a very interesting way. Enjoy the ride in the screenshots and please share your findings if you are also experimenting and exploring more unknown sides of chatGPT.