Are You the Lubricant or the Playbook?
TL;DR: A playbook captures how to do things. It cannot capture whether to do them (strategic intuition) or how to make them actually happen through people (lubrication). If your job can be distilled into a workflow, the countdown has started — regardless of your title.
Every company has three kinds of people. Two survive. One doesn’t.
The Playbook Runner
You follow a process. Maybe a sophisticated one. Maybe it took you years to learn. Doesn’t matter. Input comes in, steps happen, output comes out. No discretion. No room for the unexpected. The playbook runs, you run the playbook.
Here’s the uncomfortable part: title doesn’t save you. A senior PM who spends every day running A/B tests, pulling dashboards, writing status updates — that’s a playbook. A VP who rubber-stamps decisions based on a framework someone else built — playbook. A “Head of” who routes tickets according to a priority matrix — playbook.
The test is simple. Can someone write down what you do as a series of steps? Not roughly — actually. With enough detail that a sufficiently capable system could follow them?
If yes, you’re a playbook runner. And the clock is ticking.
The Lubricant
Some things don’t unstick themselves.
Two teams refuse to align because their leads have history nobody talks about. A deal stalls because the buyer’s real objection has nothing to do with what they said in the meeting. A product launch is blocked not by engineering, but by a political landmine between three directors who all think they own the roadmap.
The lubricant reads the room. Reads the people. Reads the thing behind the thing. They don’t follow a playbook because every situation is a new configuration of egos, incentives, and context that has never existed before and won’t exist again.
You can’t write an SOP for “get five people with conflicting incentives to nod at the same time.” You can’t flowchart “figure out that the CTO is blocking this because he’s threatened by the new hire, not because of the technical concerns he’s citing.” That’s not a process. That’s perception.
The Intuition Player
Different from the lubricant. Related, but different.
This person knows what to build. What not to build. Which market to enter. When to kill a project that metrics say is fine but gut says is dead.
Strategic intuition. Product intuition. The ability to look at twelve possible directions and pick the one that matters — not through analysis (analysis supports the decision, it doesn’t make it), but through a pattern-matching engine trained on years of context that can’t be externalized.
This is the person who says “we’re not doing that” and is right. Who kills the feature everyone loves because they see where the market is going. Who picks the weird bet that looks wrong for six months and then looks obvious.
You can be a brilliant lubricant with zero strategic intuition. Great at making things happen, no idea which things should happen. You can have extraordinary product instinct and be terrible at navigating people. Can’t get your own ideas through because you can’t read the room.
Both survive. Both are needed. They’re not the same skill.
The Distillation Test
Right now, today, the entire industry is learning to distill human work into skills and workflows. AI coding agents. AI customer service. AI data analysis. Every time someone builds a “skill” or “workflow” for an AI agent, they’re doing one thing: turning someone’s job into a playbook.
That’s what it is. You take what a person does, you write it down as steps with decision points, you hand it to a machine. Playbook-ification.
If your work can be distilled this way — congratulations, you have a clear job description. Also: the countdown has started.
Playbooks capture how. They don’t capture whether (strategic intuition) or through whom (lubrication). The first is execution. The second is judgment. The third is coordination. Execution gets automated. Judgment and coordination don’t — because they require reading contexts that change every time.
What Anthropic Is Already Showing Us
This isn’t theoretical. Anthropic’s PM org is living it.
Claude Code made 5 engineers produce like 15-20. The immediate casualty? Traditional PM work. Their response:
- Projects under 2 weeks: no PM. Engineers handle everything — including talking to legal, security, compliance. They’re mini-PMs.
- Projects over 2 weeks: PM steps in. But not to write PRDs.
- 70-80% of shipped features have no PRD. None.
What do PMs actually do now? Two things:
Direction. What to build, what to kill, where to go next. That’s the intuition play. The part no playbook captures because the answer changes based on information that hasn’t been written down yet — market shifts, competitive moves, customer signals that live in conversations, not dashboards.
Friction removal. Getting alignment between teams. Navigating conflicting priorities. Making the org machine actually move when multiple gears are grinding against each other. That’s the lubricant play.
Everything else — the project management, the status tracking, the PRD writing, the experiment running — got absorbed. By engineers. By AI. By nobody, because it turned out some of it didn’t need doing at all.
CASH — their internal AI tool — already handles A/B testing, copy changes, data pulls. Work that used to be someone’s entire job. Playbook work. Gone.
The Product-Minded Engineer
One more thing Anthropic surfaced: the product-minded engineer is now the most valuable person in the building.
Why? Because they collapse the stack. They execute (engineering), they have product intuition (what to build), and the best ones are decent lubricants too (they can navigate cross-functional friction without waiting for a PM to do it).
They’re not just following a playbook — they’re writing one, or more accurately, they’re operating in the space where no playbook exists. Building the thing and deciding what the thing should be.
If you’re an engineer who only codes what’s specced out for you: playbook. If you’re an engineer who shapes what gets built: you’re in a different category entirely.
So What Do You Do?
Audit yourself. Honestly.
Take your last week. Every meeting, every deliverable, every decision. For each one, ask: could this be written as a step-by-step process? Not “could AI do this today” — that’s the wrong question. Could it be described as a process? Could a very capable new hire follow written instructions to do what you did?
If most of your week is playbook: you have a problem. Not tomorrow. Now. The tools are already here. The distillation is already happening.
If most of your week is lubricant work or intuition calls: you’re in a different position. Not safe — nothing is safe — but different. You’re doing the work that resists playbook-ification because the inputs change every time.
The question was never “will AI take my job.” The question is: “Am I the playbook, or am I the thing the playbook can’t capture?”
一句话总结： Playbook 能抓住”怎么做”，但抓不住”做不做”（战略直觉）和”怎么让它发生”（润滑）。如果你的工作能被蒸馏成 workflow，不管 title 多高，倒计时已经开始。
每个公司里有三种人。两种能活，一种不行。
Playbook 执行者
你按流程走。也许是很复杂的流程，也许花了好几年才学会。但它还是流程。输入进来，步骤执行，输出出去。没有自由裁量，没有空间处理意外。Playbook 在跑，你在跑 playbook。
但 title 救不了你。
一个 senior PM 天天在跑 A/B test、拉数据、写周报——这是 playbook。一个 VP 按别人搭好的框架盖章做决策——playbook。一个”Head of”按优先级矩阵分配工单——还是 playbook。
测试方法很简单。有人能不能把你干的事写成一系列步骤？不是大概写——是真的写，详细到一个足够能干的系统能照着走。
能？那你是 playbook 执行者。倒计时已经开始了。
润滑剂
有些事不会自己转起来。
两个团队死活对不齐，因为两个 lead 之间有段历史，没人提。一笔交易卡住了，买方真正的顾虑跟会上说的完全不是一回事。产品发布被卡，不是因为工程问题，是因为三个 director 都觉得 roadmap 是自己的地盘，这个政治地雷没人敢踩。
润滑剂读得懂场。读得懂人。读得懂话后面那层没说出来的东西。他们不按 playbook 走，因为每次遇到的局面都是独一无二的——自尊心、利益、背景的排列组合，以前没出现过，以后也不会再出现。
你没法给”让五个利益冲突的人同时点头”写 SOP。你也没法画流程图来处理”看出来 CTO 其实不是因为技术问题反对，是因为他被新人威胁到了”。这不是流程，这是感知力。
战略直觉
跟润滑剂相关但不是一回事。
这种人知道该做什么。更重要的是，知道不该做什么。该进哪个市场。什么时候砍掉一个数据上还行但直觉告诉你已经死了的项目。
战略直觉。产品直觉。看着十二个方向，选中那个对的——不是靠分析（分析是用来支撑决策的，不是做决策的），而是靠多年积累出来的、没法外化的模式识别。
就是那个说”这个不做”然后事实证明是对的人。把所有人都喜欢的功能砍掉，因为他看到了市场要去的方向。选了个看起来很怪的赌注，六个月后所有人才恍然大悟。
你可以是极好的润滑剂，但完全没有战略直觉——特别会让事情发生，但不知道该让哪些事情发生。你也可以有极强的产品直觉，但读不懂人——想法绝对对，但推不动，因为搞不定人。
两种都能活。两种都需要。但不是同一种能力。
蒸馏测试
今天，现在，整个行业都在学怎么把人类的工作蒸馏成 skill 和 workflow。AI coding agent。AI 客服。AI 数据分析。每次有人给 AI agent 搭建一个”skill”或者”workflow”，他做的事情本质就一个：把某个人的工作变成 playbook。
就是这么回事。你把一个人干的活拆成步骤、加上判断节点、交给机器。Playbook 化。
如果你的工作能被这样蒸馏——恭喜，你的 job description 很清晰。同时：倒计时开始了。
Playbook 抓得住”怎么做”（how）。抓不住”做不做”（战略直觉）。也抓不住”怎么让它在人中间发生”（润滑）。第一个是执行，第二个是判断，第三个是协调。执行会被自动化，判断和协调不会——因为它们依赖的 context 每次都在变。
Anthropic 已经在示范了
这不是理论。Anthropic 的 PM 团队正在经历。
Claude Code 让 5 个工程师干出 15-20 个人的活。直接被挤掉的？传统 PM 的工作。他们的应对：
- 两周以下的项目：没有 PM。工程师全包——包括跟 legal、security、compliance 沟通。自己当 mini-PM。
- 两周以上才有 PM 介入。但不是来写 PRD 的。
- 70-80% 发布的功能没有 PRD。 没有。
PM 现在真正干嘛？两件事：
定方向。 做什么、砍什么、下一步往哪走。这是直觉那一层。没法被 playbook 抓住，因为答案取决于还没被写下来的信息——市场在动、竞争对手在动、客户信号活在对话里不在报表里。
去摩擦。 让团队之间对齐。处理利益冲突。让组织这台机器在齿轮互相磨的时候还能转。这是润滑剂那一层。
其他所有东西——项目管理、进度跟踪、PRD 撰写、实验执行——被吸收了。被工程师吸收、被 AI 吸收、或者被取消，因为发现有些根本不需要做。
CASH——他们内部的 AI 工具——已经在跑 A/B test、改文案、拉数据。以前是某个人全部的工作。Playbook 工作。没了。
Product-Minded 工程师
Anthropic 还暴露了一件事：product-minded 工程师现在是楼里最值钱的人。
为什么？因为他们把三层压成了一层。他们能执行（工程），有产品直觉（知道该做什么），最好的那批还是不错的润滑剂（能自己搞定跨团队摩擦，不用等 PM 来协调）。
他们不是在跑 playbook——他们在 playbook 不存在的地方操作。既造东西，又决定该造什么东西。
如果你是个工程师，只写 spec 好的代码：playbook。如果你是个工程师，能影响做什么：你已经在另一个层级了。
所以该怎么办？
审视自己。说实话。
拿你上周的工作。每个会、每个交付物、每个决策。逐一问：这能被写成步骤吗？不是”AI 今天能不能做”——这不是正确的问题。是”能不能被描述为流程”？一个很能干的新人，拿到书面指令，能不能做你做的事？
如果你一周大部分时间都是 playbook：你有问题了。不是以后，是现在。工具已经在了。蒸馏正在发生。
如果你一周大部分时间是润滑或者直觉判断：你在不同的位置。不是安全——没什么是安全的——但不同。你在做的事抗拒 playbook 化，因为输入每次都不一样。
问题从来不是”AI 会不会取代我的工作”。问题是：”我是 playbook，还是 playbook 抓不住的那个东西？”