AI collapsed my job into three roles and I had to relearn all of them
AI collapsed my job into three roles and I had to relearn all of them
AI 将我的工作压缩成了三个角色，我不得不重新学习所有内容
A message landed in my inbox a few months back. New scope, same title. Frontend, backend, and QA, folded into one role, because the team now had AI to cover the rest. I had shipped backend work for years. The frontend half of our codebase, and the QA half, I had barely touched. Reading someone else’s frontend patterns for the first time, at speed, with a deadline attached, is its own skill and I did not have it yet. I said yes anyway.
几个月前，我的收件箱里收到了一条消息。工作范围变了，但头衔没变。前端、后端和质量保证（QA）被合并成了一个角色，因为团队现在有了 AI 来覆盖其余的工作。我从事后端开发工作多年，但对于代码库中的前端和 QA 部分，我几乎从未触碰过。在有截止日期的情况下，快速阅读并理解他人的前端模式本身就是一种技能，而我当时还不具备。尽管如此，我还是答应了。
Has your role quietly grown a second and third job inside it? Not a promotion. Not new headcount. The same title, the same paycheck window, and suddenly three disciplines expected of one person because a tool now sits in the gap where two teammates used to be. I want to be honest about what that gap actually feels like from the inside, because most of what I read about this treats it as a spreadsheet problem. It is not. It is a specific, physical kind of tired that shows up on a Tuesday when you are debugging a CSS layout you have no instinct for, at 6pm, after a full day of the work you actually trained for.
你的角色是否也在悄无声息地增加了第二份甚至第三份工作？这不是升职，也不是增加人手。头衔没变，薪水没变，但因为一个工具填补了原本属于两名同事的空缺，一个人突然被要求掌握三个领域的技能。我想诚实地谈谈这种空缺在内部的真实感受，因为我读到的大多数关于此的文章都将其视为一个电子表格问题。事实并非如此。这是一种具体的、生理上的疲惫感，它会在周二下午 6 点出现——当你结束了一整天你受过专业训练的工作后，还要去调试一个你毫无直觉的 CSS 布局时，这种疲惫感尤为明显。
The tool covers the syntax, not the judgment. AI got me through the mechanical parts fast. Flexbox, a broken build step, the shape of a test file I had never written by hand. Genuinely fast. I could paste an error and get a plausible fix in under a minute. What it never gave me was the judgment for when the plausible fix was wrong for THIS codebase, and why this component was structured the odd way it was. Whether a generated test actually proved anything, or only made the red X turn green. That judgment only comes from having broken things in that specific layer before, and I had never broken anything in QA before, so I had none.
工具能覆盖语法，但不能覆盖判断力。AI 让我快速完成了机械性的工作。比如 Flexbox、损坏的构建步骤，以及我从未亲手写过的测试文件格式。确实很快。我可以在一分钟内粘贴错误并得到一个看似合理的修复方案。但它从未给我提供判断力——即判断这个看似合理的修复方案是否适合“当前”代码库，以及为什么这个组件要以这种奇怪的方式构建。它也无法判断生成的测试是否真的证明了什么，还是仅仅让红色的“X”变成了绿色的勾。这种判断力只能来自于在特定层面上曾经“搞砸过”事情的经验，而我以前从未在 QA 层面搞砸过任何事情，所以我没有这种经验。
The tool closes the syntax gap instantly. It does nothing for the judgment gap. Those are different gaps, and conflating them is where I think a lot of the current AI-productivity optimism goes wrong. What actually closed it, and what did not? Reading transcripts of AI-assisted changes other people made in the frontend and QA layers, before writing my own, closed more of the gap than writing code myself did. I learned the SHAPE of good frontend work by watching it happen, the same way I would have learned it from a senior teammate pairing with me, except the teammate was a diff instead of a person.
工具可以瞬间弥补语法差距，但对判断力差距毫无帮助。这是两种不同的差距，我认为当前许多关于 AI 生产力的乐观情绪，正是因为混淆了这两者才走入了误区。什么真正弥补了差距，什么没有？在自己写代码之前，阅读其他人利用 AI 在前端和 QA 层面上所做修改的记录，比我自己写代码更能弥补差距。我通过观察学习到了优秀前端工作的“形态”，就像我从资深同事结对编程中学到的一样，只不过这次的“同事”是一个代码差异对比（diff），而不是一个人。
What did not help was treating the AI’s first answer as correct because it compiled. A test that passes proves the test ran. It proves nothing about whether the thing it is testing actually matters. I shipped two tests in my first month that passed cleanly and tested nothing real, and only found out when a teammate in the actual QA discipline pointed it out, gently, in review.
没有帮助的做法是：仅仅因为 AI 的第一个答案能编译通过，就认为它是正确的。一个通过的测试只能证明测试运行了，它无法证明被测试的对象是否真的有意义。我在第一个月提交了两个测试，它们运行通过了，但实际上什么也没测出来。直到后来，一位真正的 QA 同事在代码审查中委婉地指出了这一点，我才发现。
The uncomfortable finding, three months in: the parts of the new scope that felt hardest were never the syntax. They were the parts where I had no scar tissue yet, no memory of a similar thing going wrong before, so I could not tell good from merely working. The part I want to say plainly: I do not think this shift is fake, and I do not think it is free. Both things are true at once. The tool genuinely lets one person cover ground that used to need three. It does not make that one person instantly as good in all three areas as three separate specialists were. Those are different claims, and the second one is the one that gets glossed over.
三个月后的一个令人不安的发现是：新工作范围中最困难的部分从来不是语法。而是那些我还没有“伤疤”的地方——没有类似事情出错的记忆，所以我无法分辨什么是“优秀”，什么是“仅仅能运行”。我想直白地说：我不认为这种转变是虚假的，也不认为它是没有代价的。这两点同时成立。该工具确实让一个人能够覆盖过去需要三个人才能完成的工作。但这并不意味着这个人能瞬间在三个领域都达到三位专业人士的水平。这是两个不同的命题，而第二个命题往往被忽略了。
If your scope has widened the same way mine did, the honest plan is not “the AI has it covered.” It is closer to what an apprentice actually does. Read more of the work in the unfamiliar layer than you write, on purpose, before you trust your own first draft there. Your turn: Which layer of the stack did AI hand you that you never actually trained for?
如果你的工作范围也像我一样扩大了，诚实的计划不是“AI 会搞定一切”。它更接近学徒的做法：在你不熟悉的领域，刻意多阅读他人的工作成果，而不是急于自己编写，直到你能够信任自己的初稿。轮到你了：AI 把技术栈中的哪一层交给了你，而你从未真正接受过相关培训？