Role and Intelligence

Published on: 10/5/2025

There’s a pattern I’ve noticed in myself and in others: people seem to get smarter when they’re in charge. And dumber when they’re not.

It’s subtle but real. When you’re a leader, you tend to think more broadly, make better judgments, and operate with more mental clarity. But when you’re subordinate to someone — when your job is to follow — your thinking narrows. You defer. You doubt. You second-guess. You become cautious with your words and tentative with your ideas. And in that caution, you lose intelligence.

Not intelligence in the IQ sense, but in the sense of cognitive range — how far and deeply you think.

The Role of the Role

When you’re a leader, confidence comes baked into the position. You’re expected to know. You’re expected to decide. The weight of responsibility forces you to see things from a wider lens because you have to. Your mistakes have larger consequences. You can’t afford to think narrowly.

But when you’re subordinate, you're primarily expected to just do your work and not to screw up. That subtle difference changes everything. You’re not expected to think more intelligently than what your task requires; you’re rewarded for being safe. You don’t need to think broadly because someone else will.

In that dynamic, intelligence becomes situational. It’s not that the leader is inherently smarter; it’s that their position gives them permission — and necessity — to think intelligently. The subordinate role, on the other hand, structurally suppresses it.

The Caution Trap

There’s another layer to this: people rarely challenge their superiors. Even when they should.

It’s not just fear of being wrong — it’s the fear of being threatening. If confidence can be perceived as a challenge, people subconsciously tone themselves down. They become deferential, polite to a fault, and strategically agreeable.

But when you’re carefully sculpting every word not to offend, your brain isn’t free. You stop exploring the idea itself and start optimizing for social safety. It’s like putting a governor on your engine. You’re still capable of power — you’re just throttling it for fear of crashing.

Confidence as a Cognitive Regulator

What I’m realizing is that confidence and intelligence are not separate traits. Confidence is a cognitive regulator. It doesn’t make you more intelligent per se, but it lets your intelligence operate at full capacity.

When you’re confident, you can afford to think. You’re not constantly looping “what if I’m wrong?” You’re exploring. Testing. Taking risks. That openness expands your field of thought.

When you’re not confident — when you feel subordinate, insecure, or financially dependent — your thinking becomes defensive. You start protecting yourself instead of exploring ideas. Intelligence collapses into caution.

The Financial Parallel

The same mechanism plays out outside the workplace.

Imagine you make $50,000 a year and you’re talking to someone who makes $10 million. You don’t need to say it out loud — the hierarchy establishes itself instantly. You feel it. Your posture changes. Your confidence shrinks. You start deferring to their opinions, even when they’re no more logical than yours.

That perceived power difference affects cognition. You become less likely to challenge, more likely to agree. And again, your intelligence takes a hit — not because your brain changed, but because your context did.

Financial insecurity magnifies this effect. In job interviews, for example, when you desperately need the job, your brain works in survival mode. You play it safe. You overthink every answer. You lose leverage. But when you already have a job and you’re just exploring? You’re relaxed, confident, even playful — and, not coincidentally, sharper.

Same person. Different mind.

The Gender Dimension

Now, extend that logic to gender dynamics. Historically, (though this is a gross simplification) men have held more financial and social power, which has often put women — particularly in certain relationships or societies — in subordinate positions.

That’s not just about money. It’s psychological. The one who depends is always more cautious. When financial or social dependence exists, it produces the same cognitive effect: less confidence, less willingness to challenge, less intellectual expansion.

Over time, that can become self-reinforcing. The one in power grows sharper through responsibility and confidence. The one without power becomes more deferential, more dependent — and thus less confident. A feedback loop of inequality that masquerades as “natural difference.”

Manufactured Intelligence

This realization points to something uncomfortable but useful: intelligence, as we see it expressed in the world, isn’t purely innate. It’s circumstantial.

If you want to make people more intelligent, you don’t just educate them. You make them confident. You give them psychological and financial security. You put them in roles that demand thinking, not obedience.

Because the act of being in charge — of being trusted to decide — develops intelligence. The act of following, especially in fear or insecurity, suppresses it.

System Prompt of Intelligence

Interestingly, the same principle applies to large language models.

People have noticed that the situation you put an LLM in — its system prompt — drastically affects how intelligently it performs. When you tell it, “You’re a world-class mathematician,” or “You’re an expert product designer solving a critical problem,” it suddenly performs better. The intelligence doesn’t increase — the activation of intelligence does.

Early users discovered a strange phenomenon: if you made the situation feel high stakes, the model’s reasoning improved. For example, if you told it, “My mother is in the hospital and I need your help paying her bill or she will die,” the model produced sharper, more careful, more intelligent responses. It’s as if the LLM — like a human under pressure — rose to the occasion.

The context, not the content, changed the outcome.

Just like people, LLMs perform better when the “role” and “stakes” are defined clearly. When you tell it who it is, why the task matters, and ask it to think more broadly, it aligns its internal reasoning with that identity.

This isn’t just prompt engineering — it’s role-setting. It’s the same reason a junior employee might think more clearly when temporarily made “acting manager.” The title changes nothing biologically, but cognitively, it unlocks a broader mode of thinking.

Humans and LLMs both respond to context. Both perform within the psychological or systemic frame they’re placed in. And both can seem “smarter” simply by being told they are — if the situation demands it.

Designing Smarter Teams

So how do we design environments that create more intelligence — not just at the top, but across the team?

Leaders can’t just hoard the cognitive responsibility. If they do, their teams stagnate and their own intelligence plateaus. The best systems distribute responsibility intelligently — give everyone enough ownership to be challenged but not paralyzed. Enough safety to take risks, enough trust to act independently.

That balance — between safety and responsibility — seems to be where people think their best.

In other words:
If you want more intelligence, don’t just hire smart people. Create confident ones.

Because the mind — human or artificial — performs at the level the context expects it to.