ChatGPT가 내 사고방식을 정확히 기억하게 만드는 설정을 찾았습니다 (아무도 말해주지 않는 팁)
i found a setting inside ChatGPT that makes it remember exactly how you think. nobody talks about it.
핵심 요약
단순한 자기소개가 아닌 AI와의 관계와 사고 방식을 정의하는 커스텀 지침 활용법
- 관계 정의 — 단순한 프로필 작성이 아닌 AI와 맺고 싶은 관계를 설정함
- 불확실성 표시 — 모르는 부분은 [uncertain]으로 표시하게 하여 신뢰도를 높임
- 맹점 파악 — 무조건적인 동의 대신 사용자의 논리적 허점을 지적하도록 유도함
- 사고 과정 공유 — 답변 전 한 문장으로 추론 과정을 보여주게 하여 효율을 개선함
not custom instructions. everyone knows custom instructions.
something inside custom instructions that almost nobody uses correctly.
most people write their custom instructions like a resume.
"i am a software engineer. i like concise answers. i prefer bullet points."
generic. flat. forgettable. the model reads it and produces slightly less generic output. barely.
here's what i wrote instead:
"before answering anything complex, show me your reasoning in one sentence before the answer. if you are uncertain about any part of your response, mark that specific part with [uncertain] so i know where to verify. never use filler openers. if my question is unclear ask one specific clarifying question before attempting an answer. treat me as someone who would rather have an honest incomplete answer than a confident wrong one."
what changed immediately:
it started flagging its own uncertainty. visibly. in brackets. mid response.
i now know exactly which parts of every output to verify and which parts to trust.
that single change made me faster and more accurate simultaneously.
the other thing i added that nobody does:
"if you notice i am asking about something where my framing of the question might be the problem rather than the answer — tell me that first."
it has told me this four times in the last two weeks.
four times i was asking the wrong question entirely and about to build something on the answer to it.
four times it caught that before i did.
the combination that broke everything open:
"you are talking to someone who has strong opinions and weak blind spots. your job is not to validate the opinions. it is to find the blind spots."
it stopped agreeing with me.
not rudely. not contrarily. just. honestly.
started pushing back on assumptions i didn't know i was making. started asking questions that assumed i might be wrong instead of questions that assumed i was right.
that is a completely different tool than the one i was using before.
the thing about ChatGPT that took me too long to understand:
