Cursor 쓰면서 코드베이스 장악력을 잃어가는 기분 드는 사람 또 있음?
Anyone else feel like they're slowly losing grip on their own codebase since using Cursor ?
핵심 요약
Cursor로 코딩 속도는 빨라졌지만, AI가 생성한 코드의 맥락을 놓치고 프로젝트 통제력을 잃어가는 것 같아 불안함을 느끼는 상황.
- 코드베이스 통제력 — AI가 작성한 코드의 맥락을 파악하기 어려워지면서 프로젝트 장악력이 떨어짐.
- 코드 리뷰 딜레마 — AI가 한 번에 여러 파일을 수정할 때 이를 일일이 검토하는 것이 생산성 향상이라는 목적과 충돌함.
- 자동화의 부작용 — AI가 아키텍처를 고려하지 않고 즉각적인 문제 해결에만 집중해 중복 코드를 생성함.
- 워크플로우 유지 — AI의 빠른 수정 속도와 개발자의 흐름을 어떻게 조화시킬지에 대한 고민이 이어짐.
Not sure if it's just me but I've been using Cursor pretty heavily for the past few months and something feels off.
The code works, shipping is faster, but I feel like I understand my own project less than I did before. Like last week I opened a file I hadn't touched in a bit and genuinely couldn't remember if I wrote that or the agent did. Spoiler : it was the agent, and it had quietly duplicated a function that already existed somewhere else.
I catch these things when I'm paying attention but that's the problem — you can't pay attention to everything when you're moving fast. The agent doesn't care about your architecture, it just solves the immediate problem in front of it.
Talked to a few people about this and the usual response is "well you should review everything the agent does" which ok, fair, but that kind of defeats the purpose no ?
Does anyone else run into this or have you found ways to actually stay on top of what the agent is doing across the whole project ?
