Chapter Four: Mental Models
The scenario decision library tells you "what to do when X happens." Mental models tell you how to think — they're the underlying operating system that lets you make sound judgments in new situations where no ready-made decision tree exists.
Model One: Your Value Isn't the Code You Write — It's the Decisions You Make
Code can be generated by AI. Any programming task that can be precisely described, AI is rapidly approaching human-level performance.
But AI can't do these things:
- "Should this system use event-driven or request-response?" — That depends on your judgment about business growth direction, the team's tech stack preferences, and what ops can realistically handle
- "Should this API be backward-compatible?" — That depends on knowing how many downstream services are calling it, and what their migration costs look like
- "What does the stakeholder actually want behind this requirement?" — What the product manager says and what they actually need are often different things. Figuring that out requires understanding people and the business
The judgment behind these decisions is the crystallization of years of experience. It doesn't live in any document, any code repository, or any screen recording of your workflow.
What this means for you: Invest your time and energy in improving decision quality, not coding speed. The latter is being caught up by AI. The former won't be.
Model Two: Information Asymmetry Is Your Only Structural Advantage
In the game between you and your company, you have one natural advantage: the company never knows what's inside your head.
Your goal isn't to hide information — hiding is easy to detect and damages your professional reputation. Your goal is to make sure the carrier of critical knowledge is always you, not a document.
The difference:
- Hiding: The company asks you, you say "I don't know" or refuse to answer → Confrontation. Risky.
- Carrier control: The company asks you, you clearly explain the operational steps and common scenarios, but that 20% of edge-case judgment requires "analysis based on the specific situation" → Cooperative. Depth under your control.
The second approach looks "professional and cooperative" from the outside — because tacit knowledge genuinely is hard to express completely in writing. That's not an excuse. That's a fact.
Model Three: Make Your Invisible Value Occasionally Visible
Your real value is often about what didn't happen:
- The bad architecture decision you blocked — without you, the team would've spent three months on a rewrite
- The technology choice you vetoed — without you, production would've crashed during peak traffic
- The security vulnerability you caught in code review — without you, a data breach makes headlines
Nobody sees these things. Their nature: when you do your job well, nothing happens. Only when you're gone does disaster prove your worth.
What this means for you: You need to strategically make these invisible contributions visible at key moments.
- In weekly/monthly reports, document "what I prevented this week," not just "what I completed"
- In architecture reviews, leave a written record of your objections and reasoning (in internal docs)
- In incident postmortems, lead the analysis — this makes you synonymous with "the person who makes the right call under pressure"
This isn't about taking credit. It's about creating evidence that your invisible value exists.
Model Four: Other People's Choices Don't Affect Your Strategy
Companies create prisoner's dilemmas on purpose — performance rankings, stack ranking with forced attrition, "AI adoption champion" awards. Some coworkers will go all-in on compliance, handing over everything. You'll watch them get short-term praise and promotion opportunities while you seem to get nothing equivalent in "rewards."
This is the moment that tests your judgment the hardest. What you need to see clearly: their advantage is one-shot — once they've handed it all over, they're back to zero. What you're protecting generates value for as long as the system is running. Full compliance doesn't mean "saved." It just means you're further back in the layoff queue.
What this means for you: Don't waste energy judging other people's choices. Don't try to convince anyone. Don't waver because of short-term comparisons. Your strategy is independent of what others do — they can only hand over what they have. They can't touch your core assets. Make your own decisions. Let time provide the answer.
Model Five: Portability Is Your Ultimate Safety Net
The deepest protection isn't "making this company unable to function without me" — because even the most irreplaceable person can get caught in a reorg.
The ultimate safety net is: "I can leave anytime."
This means your time investment should prioritize:
- Portable capabilities: Systems design thinking, deep domain understanding, cross-team communication skills, technical judgment
- Not merely becoming proficient with one company's internal tools, internal frameworks, internal processes
If you've been at a company for five years and your main skill growth is "became familiar with the company's internal systems," then your value is tied to your desk — leave the building and your market value drops sharply.
If your five years of growth look like "went from small systems to large systems, from monoliths to distributed, from executor to decision-maker," then your value is portable — it holds up wherever you take it.
Model Six: Don't Wait for Warning Signs to Start Preparing
Your company shows no signs of layoffs right now. Everything looks normal. You breathe a sigh of relief.
That sigh of relief is dangerous. It makes you do nothing during the exact window when you should be preparing.
Meta was still on a hiring spree in 2022. In 2023, they laid off 21,000 people. The window from "no warning signs" to "mass layoffs" can be just a few months.
The right decision framework isn't "are there warning signs?" but "if it happens, can I afford the consequences of being unprepared?":
| Strategy | Company IS extracting/laying off | Company is NOT extracting/laying off |
|---|---|---|
| You stay prepared | You're protected ✓ | Minimal cost (you were slightly more cautious) |
| You let your guard down | Catastrophic consequences ✗ | No cost |
The costs are asymmetric. Staying prepared costs very little — update your resume quarterly, keep 6 months of cash saved, know the difference between how and why. Letting your guard down could cost you an entire career trajectory.
This is insurance thinking. You don't buy health insurance because you "assume you'll get sick today." You buy it because you know the probability of getting sick isn't zero, and the cost of being uninsured when it happens is catastrophic.
How These Models Work Together
These six models aren't isolated tips. They're a complete thinking operating system. Internalize them, and you have the ability to make rational decisions in uncertain environments — without flipping through Zion's scenario library every time.
Worth noting: these models don't require anyone to coordinate. Each person is independently making their own best decision. But when enough people understand the logic that "handing over the why = training your own replacement," the protective choices each person makes independently will naturally converge into a collective effect — the game's equilibrium shifts. No organization needed. No alliance needed. Just the spread of information.