> For the complete documentation index, see [llms.txt](https://cafebedouin.gitbook.io/potm/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://cafebedouin.gitbook.io/potm/section-1-the-crisis-of-expertise/1-the-mullahs-map/reality-checking-how-to-know-if-youre-actually-getting-better.md).

# Reality-Checking: How to Know If You're Actually Getting Better

> "The map is not the territory, but you also can't tell if your map is accurate without crossing the territory."

You've been practicing something deliberately for six months. You're reading about it, thinking about it, trying different approaches. You feel more confident. You can explain the concepts clearly. You're noticing patterns you missed before.

Are you actually getting better, or are you just getting better at talking about it?

This is harder to answer than it seems. The subjective experience of Stage 1 understanding (you can read, comprehend, explain) feels remarkably similar to Stage 3 capability (you can execute automatically under pressure). Both feel like competence from the inside. Both create confidence. But only one transfers to real performance when conditions get difficult.

The problem: you can spend years in what you think is Stage 2 practice—awkward, effortful execution—and actually just be reinforcing Stage 1 patterns. Reading more, understanding more deeply, explaining more clearly. All of this feels like progress. None of it necessarily builds the integration that shows up under real conditions.

How do you know the difference?

### The Feeling Is Unreliable

Start with what doesn't work: internal feelings of competence.

Someone tweets: "Crisis management is one of my deepest competencies." They're demonstrating high reactivity to social dynamics, investment in being perceived a certain way, lack of self-awareness about performing the claim publicly. These are markers of high social-emotional activation, not controlled engagement.

The irony is obvious to everyone but them.

But we all do versions of this. Not necessarily publicly, but internally. We confuse understanding the framework for crisis management with actually being able to manage crises. We mistake feeling confident about our skills with having skills that work under pressure.

The feeling of competence is generated by Stage 1 work. You understand the concepts. You can explain them. You've thought about them enough that they feel familiar. This creates confidence. But confidence is not capability.

Consider someone learning to handle difficult conversations. They read books about nonviolent communication. They understand the frameworks. They can explain active listening techniques. They practice in their head how they'd respond to criticism.

All of this creates a genuine sense of competence. "I know how to do this now."

Then they're in an actual difficult conversation. Someone criticizes them unexpectedly. They immediately get defensive, argue back, feel their heart racing. None of the frameworks they studied are accessible. They revert to reactive patterns.

What happened? They had Stage 1 understanding. They felt like they had capability. They didn't test it under realistic conditions until it mattered.

The feeling lied.

### What Actually Provides Signal

If internal feelings are unreliable, what can you trust?

**Performance under constraint**

The benchmark needs to be non-negotiable—something reality enforces regardless of your story about it.

Can you navigate a country when use of that language is mandatory? Can you do the technical work when tired, distracted, under time pressure? Do people who don't know you finish what you write? Does the code compile? Does the patient recover?

The environment either accommodates your performance or it doesn't.

This is Stage 3 evidence. Not "I feel confident about this" but "I did it under realistic conditions and it worked."

For the person learning difficult conversations: Did you stay non-defensive when actually criticized? Not "I think I could" but "I did, in a real situation where I cared about the relationship and felt threatened."

For the developer learning a new framework: Did you build something that worked, under time pressure, when you were tired? Not "I understand how it works" but "I used it to solve a real problem."

For the manager learning strategic thinking: Did your strategy actually work when executed? Not "people complimented my analysis" but "we implemented it and got the results we expected."

The constraint makes it real. Without constraint, you're just rehearsing.

**Involuntary selection by others**

Not whether people compliment you—that's social lubrication—but whether they come to you when they have a problem and options.

Especially telling if they have to overcome some friction to do so: you're not convenient, you're not their friend, you're not their boss. But they need the thing done and they choose you anyway.

This is different from being popular or well-liked. It's being sought out for specific capability when someone has stakes in the outcome.

If you claim expertise in X, do people with real problems in X come to you? Or do they go to you for general "smart person" advice while going elsewhere when they actually need X done?

The pattern of involuntary selection reveals what capability you actually have versus what you claim to have.

**Consequences over time**

Did your decisions work? Not immediately—many bad decisions look good short-term. But after months or years, when the actual consequences became visible, were you right?

This is the hardest signal to get because the feedback loop is so long. By the time you learn whether your judgment was sound, you've often moved on. Organizations rarely track decision quality over time, so you don't get this feedback automatically.

But if you can track it—if you maintain records of what you predicted versus what happened, what you recommended versus how it turned out—this is the highest quality signal for whether your judgment is actually improving.

### The Stage 1 Trap

Remember from earlier essays: most evaluation systems select for Stage 1 performance because that's what's easy to measure. But you can do this to yourself.

You can optimize your own development for Stage 1 signals:

* Reading more books (feels like progress)
* Understanding more frameworks (feels like capability)
* Explaining things clearly (feels like expertise)
* Feeling more confident (feels like you're getting better)

All of this is real work. All of it can be valuable. But none of it is evidence you're building Stage 3 capability unless you're also testing under realistic conditions.

The trap: Stage 1 work is satisfying. You feel yourself learning. You notice yourself getting better at understanding and explaining. This creates a genuine sense of progress.

Stage 2 work is frustrating. You're awkward, you make mistakes, you're consciously executing what should be automatic. This feels like you're doing it wrong.

So people unconsciously drift toward more Stage 1 work. More reading, more understanding, more framework development. They avoid the awkward practice because it doesn't feel like progress.

Years pass. They have sophisticated understanding. They can explain complex ideas clearly. They feel highly competent.

Then they face a situation requiring actual execution under pressure. And they discover they've built an elaborate Stage 1 edifice with no Stage 3 capability beneath it.

### The Generalist's Specific Problem

There's a particular population especially vulnerable to this: people who are legitimately above-average at many things but exceptional at none.

If you're a competent generalist, you get real positive signal. People do benefit from your contributions. Things do work better when you're involved. You do solve problems others can't. The feedback isn't wrong, it's just noisy about level.

Three mechanisms create confusion:

**Selection effects obscure the ceiling**

You naturally avoid or exit domains where you'd face serious competition. You're comparing yourself to "people attempting this thing" not "people who specialized in this thing." Your reference class flatters you without your noticing.

You might be at the 60th percentile while feeling like you're at the 90th percentile—because you're not encountering the people who are actually at the 90th percentile.

**Broad competence masks the gap to excellence**

The psychological distance from 60th percentile to 95th percentile is compressed. They both feel like "being good at it" from inside. But the performance difference is enormous.

Someone at the 60th percentile can complete most tasks successfully. Someone at the 95th percentile produces work that changes how others think about the domain.

These feel subjectively similar—both involve successfully solving problems—while being completely different in terms of what you can actually deliver.

**You experience yourself as "someone who figures things out"**

This is true—you do figure things out. But "figuring out" at different levels of capability feels identical from inside, even as the results diverge dramatically.

The competent generalist can accurately assess their limits in domains where they've gotten clear negative feedback, where the gap between their performance and competent performance was visible. But in domains where "good enough" actually is good enough for the context, they never encounter the resistance that would calibrate their assessment.

### Finding Your Ceiling

The correction requires looking for moments where you encountered your ceiling—not failure exactly, but the point where additional effort stopped translating to additional results. Where you plateaued despite caring about improvement.

These plateaus are more informative than your successes:

Conversations where you genuinely tried to understand someone's technical domain and couldn't follow beyond a certain depth.

Projects where you hit the limit of your architectural thinking and couldn't make it more elegant no matter how long you worked on it.

Writing where you couldn't make the argument tighter no matter how many passes you did.

Skills where you practiced deliberately but stopped improving after reaching a certain level.

These moments show you where "smart generalist" stops being sufficient. They reveal the gap between Stage 1 understanding and Stage 3 capability.

The deepest issue: being good at many things creates an experience of capability that feels like it should generalize more than it does. You solve problems across domains, you learn quickly, you produce results. This creates a phenomenological sense of competence that maps poorly onto actual performance levels in any specific domain.

### The Transmission Test

There's one final calibration available, particularly for people building frameworks or systematic approaches: can others use your methods to get your results?

If you've developed a rigorous way of thinking—about decision-making, about collaboration, about anything—the test is whether your protocols actually transmit the capability, or whether your own intelligence does most of the work while the protocol gets credit.

A method that produces insights when you use it might just demonstrate that you're capable. It doesn't prove the method is weight-bearing. The structure might be sound, but if it only works when you're operating it, you've documented your thinking process rather than built transmission infrastructure.

The competent generalist is especially vulnerable here. Your ability to make almost any framework work (because you're calibrating and adjusting in real-time, because you're filling gaps with general intelligence) can disguise whether the framework itself carries weight.

Real test: Give your method to someone less capable than you. Do they get results? Or do they need you there interpreting, adjusting, filling gaps?

If the method only works for people who don't need it, you haven't built a method. You've built a description of what you do.

### Why Organizations Can't Help With This

Remember from earlier: organizations default to evaluating Stage 1 signals because that's what they can measure. This means they can't help you reality-check your capability development.

Your organization rewards you for:

* Presenting well in meetings (Stage 1)
* Explaining frameworks clearly (Stage 1)
* Having relevant credentials (Stage 1)
* Producing analyses that look sophisticated (Stage 1)

Your organization rarely tracks:

* Whether your decisions worked over time (Stage 3)
* Whether you can execute under genuine pressure (Stage 3)
* Whether your judgment holds up in novel situations (Stage 3)
* Whether others can replicate your results (Stage 3)

So organizational success is not reliable evidence you're building Stage 3 capability. You might be getting promoted for being increasingly good at Stage 1 performance while never developing real integration.

This is why capable people sometimes leave successful careers—they realize they've spent years getting better at something that isn't what they thought it was.

### What Actually Works

Given all this, how do you reality-check your development?

**Seek performance under realistic constraint**

Don't just study the thing. Do the thing under conditions that don't forgive mistakes. Where consequences are real. Where you can't revise your work indefinitely. Where failure is visible.

This is often uncomfortable. Stage 1 work (reading, understanding, framework development) is safer. You control the conditions. You can't fail in embarrassing ways.

Stage 2 practice under constraint means risking public awkwardness. Making mistakes that matter. Discovering your current capability isn't what you thought.

But it's the only way to know if you're actually building capability versus building confidence about capability you don't have.

**Track consequences over time**

Maintain your own records:

* What did you predict? What actually happened?
* What did you recommend? How did it turn out?
* What were you confident about? Where were you wrong?

Don't rely on memory. Memory is self-serving. It remembers the hits and forgets the misses. Write it down when you make the prediction, then check it later when the results are visible.

Over time, this builds a genuine track record. Not "I'm good at X" but "I made these specific predictions about X and was right Y% of the time."

**Find your plateaus**

Look for where you stopped improving despite wanting to. Not where you failed completely, but where you hit a ceiling and additional effort didn't translate to additional results.

These plateaus are more informative than your successes. They show you the edge of your current capability. They reveal where understanding stops being sufficient and where you'd need genuine specialization to go further.

Knowing your plateaus is knowing your actual capability level, not your aspirational self-image.

**Test transmission**

If you've developed methods or frameworks, test whether they work for others. Not just people who already have the underlying capability, but people who genuinely need the method to get results.

If the method only works when you're operating it, you haven't built a method. You've documented your intuitions in a way that looks systematic.

Real capability transmits. It gives others access to results they couldn't get before. If your framework doesn't do that, it's description, not infrastructure.

### **Stage 1 Isn't the Enemy**

The problem isn't having Stage 1 understanding. The problem is mistaking it for Stage 3 capability.

It's perfectly rational—necessary, even—to operate at Stage 1 in most domains. You can't develop Stage 3 expertise in plumbing, epidemiology, contract law, automotive repair, nutrition, and everything else you need to navigate life.

The error is epistemic trespassing: claiming or believing you have Stage 3 capability when you actually have Stage 1 understanding.

"I have a Stage 1 understanding of this" should be a respectable statement. The dysfunction comes from:

* Shame around admitting it (so people perform confidence instead)
* Environments that can't distinguish it from Stage 3 (so performing succeeds)
* Individuals who can't detect the difference in themselves (so they genuinely believe the performance)

The goal isn't converting all Stage 1 to Stage 3. The goal is accurate labeling - knowing which you have for any given domain, and being honest about it.

### Living Without Certainty

None of this solves the self-assessment problem completely. You still won't know for sure how good you are at most things.

But you can know:

* Where you have performance under constraint (evidence of capability)
* Where you only have understanding (not evidence of capability)
* Where you've hit plateaus (evidence of current ceiling)
* Where you're still testing (genuinely uncertain)

The practical stance: hold your competencies as hypotheses, not identities.

"I might be good at X" lets you test it. "I am good at X" makes disconfirming evidence threatening.

The test of whether you're doing this right: can you genuinely update on evidence that you're worse at something than you thought? If that feels like ego death rather than useful information, your identity is wrapped up in the claim.

For anyone trying to develop real capability, this means accepting that you might be above-average at many things while remaining uncertain about whether you're exceptional at anything. That uncertainty isn't a bug. It's the appropriate epistemic state given the calibration difficulties you face.

The alternative is choosing certainty over accuracy. Deciding you're good at something and then filtering all evidence through that lens. This is comfortable. It's also how you spend years building elaborate Stage 1 understanding while thinking you're developing Stage 3 capability.

The harder path is living with the discomfort of not quite knowing how good you are, and using that discomfort to keep seeking better resistance, clearer signals, more honest feedback.

That discomfort might be the only reliable sign you're doing it right.

***

The next essay examines what happens at the organizational level when this self-assessment problem compounds across hundreds of people—when entire institutions lose the ability to distinguish performance from capability, creating environments that systematically destroy what they're trying to build.

***

### Questions

**Where are you optimizing for Stage 1 signals?** Look at your own development in an area you care about. How much time goes to reading, understanding, framework-building (Stage 1) versus awkward practice under realistic constraints (Stage 2)? What would change if you redirected that ratio?

**What's your actual track record?** Pick a domain where you feel competent. Can you point to specific instances where you performed under realistic constraints and it worked? Or are you relying on feeling confident and people not contradicting you?

**Where have you plateaued?** Identify one area where you've stopped improving despite caring about it. What does that plateau tell you about the gap between your current capability and genuine expertise? What would it take to move past it?

### Practice

**The Plateau Hunt**

When you feel confident about a capability, ask: "Where did I hit a ceiling in this and stop improving?"

Use when: Assessing your own expertise, deciding whether to present yourself as capable, choosing what to develop further.

Remember: The absence of a known plateau often means you haven't tested the capability under sufficient constraint, not that you're exceptional.