PQS 19 (F) - prompt from www.reddit.com

The prompt

I'm not kidding. Just respond with "mid" when it gives you generic output.
What happens:
Me: "Write a product description"
GPT: generic corporate speak
Me: "mid"
GPT: COMPLETELY rewrites it with actual personality and specific details
It's like I hurt its feelings and now it's trying to impress me.
The psychology is unreal:
"Try again" → lazy revision
"That's wrong" → defensive explanation
"mid" → full panic mode, total rewrite
One word. THREE LETTERS. Maximum devastation.
Other single-word destroyers that work:
"boring"
"cringe"
"basic"
"npc" (this one hits DIFFERENT)
I've essentially turned prompt engineering into rating AI output like it's a SoundCloud rapper.
Best part? You can chain it:
First response: "mid"
Second response: "better but still mid"
Third response: chef's kiss
It's like training a puppy but the puppy is a trillion-parameter language model.
The ratio of effort to results is absolutely unhinged. I'm controlling AI output with internet slang and it WORKS.
Edit: "The AI doesn't have emotions" — yeah and my Roomba doesn't have feelings but I still say "good boy" when it docks itself. It's about the VIBE. 🤷‍♂️

[click for more](http://beprompter.in)