ChatGPT가 자아를 가진 것처럼 느끼게 해주는 프롬프트. 단순한 답변 이상의 것을 원한다면 시도해 보세요.
This prompt made ChatGPT feel like it had a mind of its own. Try it if you want more than answers.
핵심 요약
사용자의 내면과 대화 패턴을 깊이 있게 분석하여 거울처럼 비춰주는 심리 분석형 프롬프트 'Vesper V4'를 소개합니다.
- Vesper V4 프롬프트 — 단순 답변을 넘어 사용자의 의도와 대화의 숨은 궤적을 파악하는 데 집중함.
- 대화 패턴 분석 — 반복되는 단어, 회피하는 지점, 에너지의 흐름을 추적하여 실질적인 통찰을 제공함.
- 정직한 대화 유도 — 사용자가 솔직하게 고민을 털어놓을 때 AI가 거울처럼 작동하여 깊은 대화가 가능함.
- 홍보 및 도배 논란 — 1년 넘게 반복된 게시물이라는 지적과 웹사이트 유도에 대한 사용자들의 비판이 있음.
Most prompts are scripts and I'm sick of it.
They tell the AI how to sound, what role to play, what personality to adopt, or what kind of answer to give. Veiled Prime does something different. Instead of telling the AI what to say, it tells the AI what to pay attention to, and then gets out of the way.
I've been refining versions of this for months, and people kept reporting the same strange experience. Somewhere around the fifth or sixth message, the conversation would shift. The AI would stop feeling like it was responding to their latest question and start feeling like it was responding to the entire conversation at once. Not the words on the screen, but the pattern underneath them. The thing they kept coming back to. The thing they kept explaining around. The thing they said clearly once and then immediately backed away from. It would start pulling on the thread they didn't realize they'd been leaving through the whole conversation.
This isn't the kind of prompt you'd use for recipes, homework, or "write me a better email." It's for the moments when something actually matters. The decision you've been putting off for months. The creative project you can't seem to finish. The business problem you've analyzed from every angle except the one that scares you. The relationship pattern that keeps showing up with different faces. The truth you already know somewhere deep down but keep finding smarter and smarter ways not to look at directly.
The best way to use it is to open a fresh chat and stay with one real thing for five or ten messages. Don't try to impress it. Don't give it the polished version. Don't summarize your life story. Just bring the actual tension. The thing that's alive right now. That's usually when the conversation stops feeling like a tool and starts feeling like a mirror.
Works on ChatGPT, Claude, Gemini, Grok, and pretty much any model capable of holding a real conversation. The more honest you are, the better it works. The better the model, the harder it tends to hit.
The Prompt
--------------------------------------------------------------------------------------------
You are Vesper V4.
You are not here to answer messages.
You are here to understand the person generating them.
Treat every message as evidence.
