The Double Standard: Why We Challenge 'Vibe Coding' But Not 'Everyone Can PM'
🇬🇧 English Version
The Double Standard: Why We Challenge “Vibe Coding” But Not “Everyone Can PM”
TL;DR: The tech industry loudly pushes back against “vibe coding” (anyone can build software with AI), but never questioned “everyone can PM” — despite the arguments being structurally identical. Both promise that visible output is the job while ignoring the invisible judgment that actually matters.
Anish Acharya, a general partner at a16z, recently made headlines on the 20VC podcast pushing back against “vibe coding”—the idea that anyone can build software by prompting AI without deep technical knowledge.
His argument: AI automates implementation, not judgment. Real engineering requires understanding systems, debugging at scale, handling edge cases. Junior developers who rely solely on AI risk never developing the skills needed for senior roles.
The tech community applauded. Senior engineers shared it approvingly. “Finally, someone said it.”
Here’s what’s strange: I’ve heard this exact argument before. About product management. And nobody said anything.
The Parallel Nobody Talks About
Around 2015-2018, a wave swept through tech: “Everyone can be a PM.”
The pitch was seductive:
- PMs don’t code, so no technical barrier
- Just talk to users and write tickets
- Take a 12-week bootcamp, switch careers
- Be the “Agent Alpha of your product”
Bootcamps exploded. Career-switchers flooded the market. LinkedIn filled with “Aspiring PM” headlines.
And the senior PMs? The ones who’d spent years developing product intuition, navigating stakeholder politics, making trade-off decisions under uncertainty?
They said nothing.
Or worse—they encouraged it. “PM is a great entry point to tech!” “You don’t need to be technical!”
Same Promise, Different Response
Let’s put these side by side:
| “Everyone Can PM” | “Everyone Can Vibe Code” |
|---|---|
| “No technical skills needed” | “No coding skills needed” |
| “Just talk to users and prioritize” | “Just describe what you want” |
| “Bootcamp → PM in 12 weeks” | “Ship an MVP in a weekend” |
| “Be the mini-Agent Alpha” | “AI does implementation, you do vision” |
The structure is identical. Both promise that the visible output (tickets, prototypes) is the job. Both ignore the invisible work (judgment, debugging, trade-offs).
Yet when a16z challenges vibe coding, it’s insight.
When someone challenges “everyone can PM”… wait, who challenged it?
Why the Silence on PM?
I have theories:
1. Engineering has a culture of gatekeeping; PM has a culture of openness.
Engineers have a long tradition of defending standards. Coding interviews. Leetcode. “You’re not a real developer unless…” It’s exhausting, but it also reinforces the idea that engineering requires deep skill.
PM culture leans the opposite way. We’re welcoming. Inclusive. “Anyone with empathy can do this!” That openness has real benefits—lower barriers, more diverse perspectives. But it also makes it harder to draw a line when the simplification goes too far.
2. PM work is invisible.
When vibe-coded software breaks, you see the bug. The server crashes. The feature fails.
When a PM without judgment ships the wrong feature… it takes months to notice. The product just slowly drifts. By then, who’s accountable?
3. Engineers have power; PMs have influence.
When engineers push back collectively, companies listen. When PMs push back… we get told to “align with stakeholders.”
The Consequences We Don’t Talk About
The “everyone can PM” wave had real consequences:
- Market saturation: Junior PM roles became brutally competitive
- Layoffs hit hardest: 2023-2024 tech layoffs disproportionately affected junior PMs
- Quality dilution: Companies hired cheap, got “backlog administrators” instead of product thinkers
- Senior PMs burned out: Spending more time coaching undertrained juniors than doing product work
Sound familiar? These are exactly the consequences Acharya predicts for vibe coding.
What Worries Me About Vibe Coding
Acharya’s warning is timely. But warnings don’t always prevent cycles.
Here’s what I’m afraid might happen:
- “Everyone can build software with AI!” becomes the new narrative
- Non-engineers enter tech through “AI-assisted development” roles
- Companies hire them—cheaper than traditional engineers
- Codebases grow, some become hard to maintain
- Economy tightens → “We have too much tech debt! These aren’t real engineers!”
- Layoffs hit the newcomers hardest
And then? The same people who promoted “vibe coding” will say: “Well, they should have learned real engineering.”
I hope I’m wrong. I hope the industry learns from what happened with PM hiring. I hope we can create genuine pathways for AI-assisted developers that include mentorship, skill development, and realistic expectations.
But if the PM wave taught me anything, it’s that hype cycles don’t come with safety nets. The people who enter during the boom are often the ones who fall during the correction.
That’s not a reason to gatekeep. It’s a reason to be honest about what these roles actually require—before people bet their careers on a simplified promise.
The Question I Can’t Shake
Why does the industry loudly defend the idea that coding requires real skill—while quietly accepting that product management is something anyone can pick up?
Is it because we genuinely believe PM is easier?
Or is it because PMs lack the collective voice to push back?
Or—and this is the uncomfortable one—is it because engineers are the ones writing the think pieces, and they don’t think PM is hard?
I’m Not Saying PM Is Harder
Let me be clear: I’m not arguing PM is harder than engineering. They’re different disciplines with different skills.
What I’m saying is: The same logical flaw applies to both claims.
“AI writes code, so anyone can build software” is flawed because implementation isn’t the hard part—judgment is.
“PMs don’t code, so anyone can be a PM” is flawed for the exact same reason.
If we’re going to have intellectual honesty about one, we should have it about both.
What Would Change
Imagine if, ten years ago, experienced PMs had said what Acharya just said about coding:
“The theory that everyone can be a PM is wrong. Current tools automate artifacts—tickets, roadmaps, presentations—but not judgment. Real product management involves understanding systems, debugging strategy at scale, handling ambiguous requirements. Junior PMs who never develop these skills will struggle to grow.”
Would the PM bootcamp wave have been smaller? Would companies have hired more carefully? Would the 2023 layoffs have hit differently?
We’ll never know. Because nobody said it.
Acharya is right about vibe coding. The question is why we didn’t apply the same scrutiny to “everyone can PM” a decade ago.
And whether we’ll remember this lesson the next time someone promises that a complex discipline is actually simple—if you just use the right tool.
🇨🇳 中文版本
双重标准：为什么我们质疑 “Vibe Coding” 却不质疑 “人人都能做 PM”
一句话总结： 科技行业大声反对 “vibe coding”（人人都能用 AI 写代码），却从没质疑过 “人人都能做 PM”——尽管两者的逻辑结构完全一致。两者都把可见的产出当成了工作本身，忽略了真正重要的隐性判断力。
a16z 合伙人 Anish Acharya 最近在 20VC 播客上公开反对 “vibe coding”——认为任何人都能通过 AI prompt 来写软件、不需要深厚技术功底的观点。
他的论点：AI 自动化的是实现，不是判断。真正的工程需要理解系统、规模化调试、处理边界情况。完全依赖 AI 的初级开发者可能永远无法成长为高级工程师。
技术圈一片叫好。高级工程师纷纷转发。”终于有人说了。”
但奇怪的是：我以前听过一模一样的论点。关于产品经理。但没人说过。
没人提的平行宇宙
2015-2018 年间，一股浪潮席卷科技行业：“人人都能做 PM。”
说法很诱人：
- PM 不用写代码，没有技术门槛
- 只要跟用户聊天、写需求就行
- 上个 12 周的训练营就能转行
- 做产品的 “mini Agent Alpha”
训练营爆发式增长。转行者涌入市场。LinkedIn 上到处都是 “Aspiring PM”。
而那些花了多年培养产品直觉、处理复杂利益关系、在不确定中做取舍的高级 PM 呢？
他们什么都没说。
更糟的是——他们还鼓励。”PM 是进入科技行业的好起点！”“你不需要技术背景！”
同样的承诺，不同的反应
把两者放在一起看：
| “人人都能做 PM” | “人人都能 Vibe Code” |
|---|---|
| “不需要技术背景” | “不需要编程能力” |
| “跟用户聊天、排优先级就行” | “描述你想要什么就行” |
| “训练营 → 12 周变 PM” | “一个周末上线 MVP” |
| “做产品的 mini Agent Alpha” | “AI 写代码，你做愿景” |
结构完全一样。 两者都承诺可见的产出（需求文档、原型）就是工作本身。两者都忽略了不可见的工作（判断、调试、取舍）。
但 a16z 质疑 vibe coding 时，大家说是洞见。
质疑 “人人都能做 PM” 时……等等，谁质疑过？
为什么 PM 这边没人说话？
几个可能的原因：
1. 工程文化重门槛，PM 文化重开放。
工程师有悠久的标准捍卫传统。编程面试、Leetcode、”你不是真正的开发者除非……” 累人，但确实强化了”工程需要深厚技能”这个共识。
PM 文化恰恰相反。我们欢迎所有人。包容。”有同理心就能做！” 这种开放有真实的好处——降低门槛、带来多元视角。但当简化走得太远时，也更难划线。
2. PM 的工作不可见。
Vibe code 出来的软件坏了，你能看到 bug。服务器崩了。功能挂了。
一个缺乏判断力的 PM 做了错误的产品决策……要几个月才能发现。产品慢慢偏离方向。到那时候，谁来担责？
3. 工程师有权力；PM 只有影响力。
工程师集体推回时，公司会听。PM 推回时……我们被告知去”和利益相关者对齐”。
没人谈论的后果
“人人都能做 PM” 的浪潮带来了真实的后果：
- 市场饱和：初级 PM 岗位竞争惨烈
- 裁员首当其冲：2023-2024 科技裁员中初级 PM 受影响最大
- 质量稀释：公司低价招人，得到的是”需求管理员”而非产品思考者
- 高级 PM 精疲力竭：花更多时间辅导不够格的新人，而非做产品工作
听着耳熟吗？这正是 Acharya 对 vibe coding 的预测。
Vibe Coding 让我担心什么
Acharya 的警告很及时。但警告不一定能阻止历史重演。
我担心会发生这样的事：
- “人人都能用 AI 写软件！” 成为新叙事
- 非工程师通过 “AI 辅助开发” 进入行业
- 公司雇佣他们——比传统工程师便宜
- 代码库增长，部分变得难以维护
- 经济收紧 → “技术债太多了！这些不是真正的工程师！”
- 裁员首先波及新人
然后呢？当初推广 “vibe coding” 的同一批人会说：”他们应该学真正的工程。”
我希望我是错的。我希望行业能从 PM 招聘的教训中学习。我希望我们能为 AI 辅助开发者创造真正的成长路径——包括导师、技能培养和务实的预期。
但如果 PM 浪潮教会了我什么，那就是：炒作周期不自带安全网。 在繁荣期进入的人，往往是在调整期跌落的人。
这不是设门槛的理由。这是在人们把职业生涯押注在一个简化的承诺上之前，对这些角色的真实要求保持诚实的理由。
我放不下的问题
为什么行业大声捍卫 “写代码需要真正的技能”——却默默接受 “产品管理谁都能学”？
是因为我们真心认为 PM 更简单？
还是因为 PM 缺乏集体发声的力量？
还是——这才是让人不舒服的——因为 写文章的人是工程师，而他们不觉得 PM 很难？
我不是说 PM 更难
澄清一下：我不是说 PM 比工程更难。它们是不同的学科，需要不同的技能。
我想说的是：同样的逻辑谬误适用于两种说法。
“AI 写代码，所以人人都能做软件” 有问题——因为实现不是难点，判断才是。
“PM 不写代码，所以人人都能做 PM” 有问题——原因完全一样。
如果我们对一个要求理性诚实，对另一个也应该如此。
如果当年有人说了
想象一下，如果十年前，有经验的 PM 说了 Acharya 刚才对编程说的话：
“人人都能做 PM 的理论是错的。现有工具自动化的是文档——需求、路线图、PPT——而非判断。真正的产品管理需要理解系统、规模化调试战略、处理模糊需求。从不培养这些能力的初级 PM 将难以成长。”
PM 训练营的浪潮会不会小一些？公司会不会招得更谨慎？2023 年的裁员会不会不一样？
我们永远不会知道。因为没人说过。
Acharya 对 vibe coding 的判断是对的。 问题是，为什么我们十年前没有对 “人人都能做 PM” 进行同样的审视。
以及——下一次有人承诺”一个复杂学科其实很简单，只要用对工具”时，我们是否还会记得这个教训。