High Agency: The One Trait That Actually Matters

I keep coming back to the same idea: high agency.

It’s become a trendy term lately, showing up everywhere. Eric Weinstein coined it back in 2016 on Tim Ferriss’s podcast, describing people who hear “that’s impossible” and immediately start figuring out how to get around it. But I first encountered it through Shreyas Doshi’s frameworks, which made the concept concrete and actionable.

Before the phrase caught on, I had a simpler way of putting it: I hire for skill and attitude. Skill you can test. Attitude is what makes the difference.

High agency is just a better name for that “attitude” part. It shows up in how I think about hiring, team culture, and what separates good engineers from great ones. For the kind of roles I typically hire (senior engineers, tech leads, people who need to operate with autonomy), it’s the closest I’ve found to a single trait that predicts whether someone will thrive or struggle.

What High Agency Actually Means

Shreyas Doshi has a great framework for this. He contrasts Low Agency Bob with High Agency Alice. Bob says “I cannot get this done because…” and lists external blockers: legal will block it, we don’t have resources, the strategy is unclear, my boss doesn’t have time. Alice says “I will do my best to get this done because…” and focuses on what she controls: it’s important, I can overcome obstacles, I’ll learn from this, I’m persistent.

Same situation. Completely different framing.

High agency is the tendency to act rather than wait. To see a problem and fix it instead of flagging it for someone else. To find a way when the obvious path is blocked.

It’s not about ignoring constraints or being reckless. It’s about treating obstacles as problems to solve, not reasons to stop.

Low agency sounds like: “That’s not my area.” “I’m blocked, waiting on X.” “Nobody told me to do that.” “That’s how it’s always been done.”

High agency sounds like: “I noticed this was broken, so I fixed it.” “I couldn’t reach them, so I found another way.” “The spec was unclear, so I made a call and documented why.”

The difference isn’t intelligence or skill. It’s a mindset. A belief that you can influence outcomes, combined with the initiative to actually do it.

Andrej Karpathy put it well: “Agency > Intelligence. I had this intuitively wrong for decades, I think due to a pervasive cultural veneration of intelligence, various entertainment/media, obsession with IQ etc. Agency is significantly more powerful and significantly more scarce.”

He’s right. We obsess over hiring the smartest people in the room. But smart people who wait for permission accomplish less than average people who just start.

Why It Matters More Than Process

Process exists to coordinate people who need coordination. The more high-agency people you have, the less process you need.

I’ve seen teams with perfect Scrum ceremonies ship nothing of value. I’ve seen two engineers with a shared understanding build something remarkable in a week. The difference wasn’t methodology. It was the people.

High-agency engineers don’t wait for the sprint to end to raise a concern. They don’t need a ticket to fix something obviously broken. They don’t treat the backlog as the complete set of things worth doing.

This doesn’t mean chaos. High agency without alignment is just people running in different directions. But given clear goals and context, high-agency people will find ways to hit those goals that no process document could anticipate.

What Changes When You Have More of Them

The more high-agency people you have, the more your organization can do with the same headcount.

This sounds obvious, but the implications run deep. Traditional org structures assume most people need coordination. Managers translate strategy into tasks. Process ensures consistency. Approval layers catch mistakes. Each layer exists because someone, somewhere, didn’t trust the layer below to make good decisions.

High-agency people break this model. They self-coordinate. They make good decisions without approval gates. They act on context, not instructions. The coordination overhead that justifies middle management starts to shrink.

This doesn’t mean managers become useless. It means their role shifts. The valuable work becomes:

• Providing context, not assigning tasks. High-agency people need to understand the why, not be told the what.
• Removing obstacles, not approving decisions. Clear the path, then get out of the way.
• Coaching and developing, not supervising. Help people grow, don’t check their work.
• Handling cross-team dependencies that individuals can’t resolve alone.
• Shielding the team from organizational noise so they can focus.

Managers who can’t make this shift become the “middle layers that exist to say no.” High-agency people route around them, or leave.

The organizations that figure this out achieve more. Not through heroics or overtime, but through leverage. Every layer you remove, every approval gate you eliminate, every decision you push down, compounds into speed and adaptability. This is why small, high-agency teams often outperform much larger organizations. It’s not just talent density. It’s the absence of coordination tax.

The constraint becomes alignment, not control. Get that right, and high-agency people become a multiplier. Get it wrong, and you have talented individuals pulling in different directions.

Beyond Code: Product-Minded Engineers

High agency isn’t just about shipping faster. It’s about understanding what’s worth shipping in the first place. Product-mindedness tells you what matters; agency ensures it happens. They’re different traits, but high-agency engineers tend to develop product intuition because they care about outcomes, not just output.

The best engineers I’ve worked with don’t just write code. They understand the problem. They ask why before how. They push back on requirements that don’t make sense and propose alternatives that actually solve the user’s need.

This is what separates a feature factory from a product team. Feature factories optimize for output: tickets closed, story points burned, releases shipped. Product teams optimize for outcomes: did we actually solve the problem? Did the metric move? Did users benefit?

High-agency engineers naturally gravitate toward outcomes. They’re not satisfied with “I built what was asked.” They want to know if it worked. They follow up. They look at the data. They talk to users. When something isn’t landing, they don’t wait for the next sprint planning to raise it.

This connects to how we think about product at SeQura: solve real problems, not feature checklists. That only works if you have engineers who care about the problem, not just the implementation.

You can have the best product strategy in the world. If your engineers are waiting to be told exactly what to build, you’ll still end up with mediocre products. But give a high-agency engineer a real problem and context, and they’ll often find a better solution than whatever was in the spec.

How to Spot It in Hiring

You can’t ask “Are you high agency?” and expect a useful answer. Everyone says yes. You have to dig for evidence.

I look for stories where the candidate did something they weren’t asked to do. Not heroics or rule-breaking, just initiative. “I noticed X, so I did Y.” The best answers are mundane. Fixing a flaky test nobody assigned. Reaching out to a customer directly to understand a bug. Writing documentation because onboarding was painful.

Red flags: every accomplishment framed as a team effort with no individual ownership. Blaming failures entirely on external factors. Detailed knowledge of process but vague answers about outcomes.

Green flags: concrete examples of autonomous decisions. Comfort with ambiguity. Reflection on mistakes without excessive defensiveness. A track record of figuring things out.

One question I like: “Tell me about a time you did something important that wasn’t your job.” The answer tells you a lot. Some people light up. Others struggle to find an example. Both responses are informative.

How Culture Enables or Kills It

Here’s the uncomfortable part: most organizations systematically destroy high agency.

Every approval layer says “your judgment isn’t trusted.” Every rigid process says “follow the rules, don’t think.” Every time someone gets burned for taking initiative, everyone else learns to keep their head down.

High-agency people don’t stay in low-agency environments. They leave, or worse, they adapt. They learn that initiative isn’t rewarded, so they stop showing it. You’ve taken your best people and trained them to be average.

Building a high-agency culture means:

Tolerating mistakes made in good faith. If someone makes a reasonable call that turns out wrong, you don’t punish them. You discuss what you’d do differently and move on. The moment people fear being blamed, they stop deciding.

Pushing decisions down. The person closest to the problem usually has the best context. Let them decide. “I trust your judgment” has to be more than words.

Rewarding outcomes, not activity. Story points completed means nothing. Problems solved for users means everything. When you measure the right things, high-agency behavior follows.

Making context available. People can’t make good autonomous decisions without understanding the bigger picture. Share strategy, share constraints, share why things matter. Opacity kills agency.

Does Everyone Need High Agency?

No. And that’s fine.

You need high agency at the edges, where people see problems first. The person on call at 3am needs to make decisions, not file tickets. You need it in leadership, where waiting for perfect information paralyzes everyone else. And you need a baseline of ownership everywhere: “this is my problem” about your immediate work.

But some roles benefit from consistent execution over creative problem-solving. A compliance analyst who follows the process exactly is doing their job well. That’s not low agency. That’s judgment about when structure matters.

Here’s the distinction: ownership is caring, agency is acting. The bar for both varies by role, but it’s never zero.

Our product manifesto sets the floor for ownership: you own what you build, you speak up when you see a problem, you care about outcomes not just delivery. That’s baseline ownership, not high agency.

High agency adds initiative. You don’t just care about the problem, you act without waiting. You don’t just flag blockers, you find another way. For senior roles, leadership positions, and anyone operating with autonomy, that’s where I filter hard.

A team with 30% high-agency operators and 70% solid executors can be incredibly effective, if the high-agency people are in the right spots. The 70% aren’t passive. They just prefer clarity over ambiguity, and that’s fine. What’s not fine is hiring people with no initiative at all.

What kills big teams isn’t lack of universal agency. It’s agency concentrated only at the top. Middle layers that exist to say no. Punishing mistakes more than inaction. The question isn’t “is everyone high agency?” It’s “when agency shows up, does the organization reward it or crush it?”

Leading Mixed Teams

So if not everyone needs to be high agency, how do you actually lead a mixed team?

Here’s something high-agency leaders often get wrong: they assume everyone wants autonomy.

Scott Stevenson nailed this: “A lot of high agency people who would make great executives would make terrible middle managers, because they don’t understand the desire to be led that many people have—because they have never experienced this themselves.”

This hit home for me. Early in my leadership journey, I’d give people context and space, then get frustrated when they wanted more direction. I thought I was empowering them. They thought I was abandoning them.

The truth is, ambiguity is stressful for many people. Some genuinely thrive with clear direction, defined scope, and predictable structure. They want to know exactly what success looks like and then execute well. That’s not a flaw. It’s a different mode of working, and it’s valuable.

The mistake is treating everyone the same. High-agency people suffocate under micromanagement. Lower-agency people flounder without structure. A good leader reads what each person needs.

For high-agency people: Give them the problem, not the solution. Share context and constraints, then get out of the way. Check in on outcomes, not activity. Let them surprise you.

For people who want more direction: Be specific about what good looks like. Break work into clear milestones. Check in more frequently, not to control, but to provide the clarity they need to feel confident. Gradually expand scope as they build comfort.

For the team overall: Create explicit norms. When is autonomy expected? When should people escalate? What decisions can be made alone vs. need input? High-agency people assume “figure it out” is the default. Others assume “check first” is the default. Make the actual expectation clear.

Sometimes that means pairing a high-agency person with someone who’s great at execution. Sometimes it means putting guardrails on your most autonomous people so they don’t run too far ahead. The goal is getting the best from each person, not turning everyone into the same type of operator.

The Trade-offs

The downsides are real.

High-agency people can be exhausting. They push back. They question decisions. They sometimes act before checking with everyone who wants to be checked with. If you need conformity, they’re a bad fit.

They also require trust and context to function well. In environments with genuinely limited information (regulated industries, early security-sensitive work), high agency can be dangerous. Sometimes you actually need people to follow the process exactly.

And a team of entirely high-agency people can have coordination problems. Someone has to care about the boring stuff. Not every task needs reinvention.

But in my experience, these trade-offs are worth it. The alternative, a team that waits for permission, that treats every problem as someone else’s responsibility, that optimizes for process compliance over outcomes, is far worse.

Skill and Will

There’s a phrase I come back to: skill and will are everything, but assuming skill is a given, will is what makes the difference.

When I hire, I assume we can verify skill. Technical interviews, take-home assignments, reference checks. It’s not trivial, but it’s solvable. You can test for aptitude.

What’s harder to test is attitude. The will to act. The drive to take ownership without being asked. That’s high agency.

Shreyas Doshi maps this as a 2x2. High talent plus high agency gives you game changers. High talent plus low agency gives you frustrated geniuses: brilliant people who can’t get things done because they’re stuck waiting for perfect conditions. Low talent plus high agency gives you go-getters: they’ll figure it out, learn what they need, and deliver. Low talent plus low agency gives you cogs-in-the-wheel.

The insight: go-getters often outperform frustrated geniuses. Someone with moderate talent who acts beats someone with exceptional talent who waits. That’s why agency matters more than raw ability.

Early in my career, I obsessed over technical skills, specific experience, culture fit (whatever that meant). I hired people who interviewed well and looked good on paper. Results were mixed.

Now I care about three things: Can they do the work? Will they do the work without being managed? Do they make the people around them better?

The second question is high agency. And it’s the one I’ve found hardest to train, which makes it the one most worth filtering for.

You can teach someone a new language, a new framework, a new domain. You can’t easily teach someone to take ownership. That comes from somewhere deeper.

So when I’m building a team, when I’m deciding who to bet on, high agency is the filter I keep coming back to. Skill determines what you can do. Agency determines what you will do. In the long run, I bet on the latter.