My bottleneck moved from writing code to reviewing what the agents write. How do you review without just rubber-stamping?
A developer noted that their bottleneck shifted from writing code to reviewing agent-generated code, leading to concerns about rubber-stamping approvals. To manage this, they keep tasks small, request agent notes, and review tests first. They also use a review agent to flag messy code, which adds to the review load. The developer is seeking advice on balancing careful review with trusting agent output, questioning whether to read every line or trust certain changes more.
时间与来源
时间显示为 UTC
显示时区：UTC
本地时区尚不可用，暂时显示 UTC。
发布当时偏移：UTC+02026年10月29日 19:05 UTC
收录当时偏移：UTC+02026年10月29日 22:00 UTC
- 发布
- 2026年10月29日 19:05
- 收录
- 2026年10月29日 22:00
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
档位是按信源手工设定的编辑判断，不是逐条打分。
Since I started running a few agents in parallel I can get several changes ready in the time it used to take me to write one, and that part feels great. But now most of my day is reading diffs, and I've caught myself skimming a big one and approving it just because the tests passed. A few things help so far. I keep each task small so the diff fits on one screen, I ask the agent for a short note on what it changed and why before I read the code, and I read the tests first to see if they actually check the behavior I asked for. A review agent also flags old messy code nearby, which is nice, but then I have even more to read. I still feel like I'm the slow part and I'm not sure where the line is between careful review and just trusting it. How do you all handle this? Do you read every line, or do you trust some kinds of changes more than others?