Agency Is the Only Skill That Compounds
Agency, the trained habit of acting without waiting for permission, is the meta-skill that multiplies every other skill, and it matters more as AI makes raw execution cheap.
Two people start the same job on the same Monday. Same degree, same test scores, roughly the same raw ability. Five years later one of them is running a team and shipping things the company didn’t know it needed, and the other is still waiting to be told what to work on. The gap did not come from talent. It came from a single habit, invisible on any resume, that one of them practiced ten thousand times and the other never did.
That habit is agency. And I want to make a stronger claim than the usual pep talk about being proactive. Of all the skills you can build, agency is the only one that compounds, because it is the skill that decides how much of every other skill you ever actually use.
What agency actually is (and what it isn’t)
Let me define it precisely, because the word gets used loosely. Agency is the trained disposition to act on the world without waiting for permission. Three parts, and all three have to be there.
- Initiative. You start things. You see a gap and move toward it before anyone assigns it to you. Not recklessly, but you treat the absence of instructions as a green light rather than a red one. Most people treat it as a red one, which is exactly why initiative is rarer and more valuable than it should be.
- Ownership. Once you start, the outcome is yours. You don’t hand off the hard part and call it someone else’s problem when it breaks. Ownership is what turns a one-time burst of initiative into a track record, because you stay with the thing until it actually works.
- A bias to try. When you’re unsure, you lean toward the experiment over the deliberation. You’d rather learn from a small real attempt than argue about a hypothetical in a meeting. This is the muscle that beats analysis paralysis, and it is the one most people have let go soft.
Notice what agency is not. It is not talent. A brilliant person who waits to be told what to do converts almost none of that brilliance into results. It is not confidence either, though people confuse the two constantly. Confidence is a feeling about yourself. Agency is a behavior toward the world, and you can practice the behavior while the feeling is still shaky. In fact that’s the whole trick. You act first, and the confidence arrives later as a receipt.
Why agency compounds when other skills only add
Here’s the mechanism, and it’s worth slowing down for, because this is the part that makes agency different in kind, not just in degree.
Most skills add. Learn a second language and you can now speak a second language. Useful, linear, done. The skill sits there as a fixed asset until you use it. Agency is different. Every act of agency changes what you believe is possible, and what you believe is possible sets the ceiling on what you’ll even attempt next.
Walk through one real loop. You ship a small thing nobody asked for. A tiny internal tool, a rewritten page, a fix you weren’t assigned. It works. Two things happen. First, the world responds: someone uses it, thanks you, points you at a bigger problem. Second, and more importantly, your internal map updates. The category of “things I’m allowed to just do” gets slightly larger. So next time the bar you’ll clear on your own is a little higher. You attempt the bigger thing. It works, or it half-works and you fix it. The map expands again.
“Every act of agency expands what you believe you’re allowed to attempt, and that expanded belief is what makes the next, bigger act possible. Skill adds. Agency multiplies the skill you’ll ever actually deploy.
”
That is a compounding loop, in the exact mathematical sense. Each act raises the baseline the next act starts from. Meanwhile the person who waits is running the loop in reverse. Every time they wait to be told, the category of “things I’m allowed to just do” shrinks a little, the permission they require grows, and their real capability, whatever their raw talent, quietly shrinks to fit the smaller box. Same starting skills, opposite exponents. Give that five years and the two lives are barely comparable.
This is why I’ll call agency the meta-skill. It doesn’t sit next to your other skills. It sits above them, deciding how much of each one ever leaves the shelf.
Why agency gets more valuable exactly as execution gets cheap
Now the part that makes this urgent rather than merely nice.
For most of history, execution was the bottleneck. Knowing what to do was the easy half. Actually building the thing, writing the code, making the design, drafting the document, cost weeks of skilled labor. So the scarce, well-paid ability was execution, and agency was a bonus on top.
That ratio is inverting fast. When a capable person with the right tools can generate a working draft of almost anything in minutes, the cost of doing collapses toward zero. And when the cost of doing collapses, the scarce thing is no longer the doing. It’s the deciding. Deciding what is worth doing, and then actually starting, becomes the entire job.
Read that again, because it is the crux. In a world where the machine will execute nearly anything you point it at, the value moves entirely to the person who points. Pointing is an act of agency. It is choosing a direction with no one to confirm it’s the right one, and committing before the outcome is certain. The tools amplify whoever is driving, and they do nothing at all for the person still waiting to be handed a task, because the task was the part that just got cheap.
This connects directly to two other ideas worth chasing. One is taste: once you decide to act, taste is what tells you which of a thousand generated options is actually good. The other is the company of one: agency plus modern tools is what lets a single person now do what used to require a whole team, because the team’s execution work is exactly the part that got automated. Agency is the ignition for both.
Agency without judgment is just recklessness
Time for the honest caveat, because the failure mode is real and I’ve watched it burn people.
Agency detached from judgment isn’t a virtue. It’s a hazard. The person who ships constantly without ever stopping to ask whether it’s the right thing, who overrides people who know more, who mistakes motion for progress, is not exercising agency in the sense that matters. They’re just loud. Raw initiative without ownership of the consequences is how you become the colleague everyone quietly works around.
The distinction is between reckless agency and learned agency. Reckless agency acts to feel decisive. Learned agency acts to find out something true, and stays humble about what the outcome teaches. The difference shows up in one behavior: what you do when the attempt fails. The reckless version blames the world and moves on unchanged. The learned version reads the failure, updates, and tries a smarter version. That feedback loop is the entire safety mechanism. Agency without it is a car with an accelerator and no steering.
So the goal is not “act more.” It’s “act more, on things that matter, in a way that teaches you something whether or not they work.” The reps build judgment at the same time they build nerve, as long as you actually look at the results.
How to train agency (it’s a muscle, not a trait)
The best news in all of this: agency is trainable. People talk about it like a personality type you either have or you don’t. It isn’t. It’s a disposition built from reps, and the reps are small. Here is how to run them.
Shrink the thing until you can start today
The enemy of agency is the project so big that starting feels like a commitment you’re not ready to make. So shrink it. Not “launch the product,” but “sketch one screen tonight.” Not “learn the language,” but “write ten lines that run.” A rep you can start in the next hour is worth more than a grand plan you’ll begin someday. Momentum is the only thing that compounds, and momentum starts small.
Cut the permission you're waiting for
Make a habit of asking one question: whose approval am I actually waiting on, and do I truly need it? Half the time the answer is nobody, and you were waiting out of habit. The reversible decisions, the ones you can undo cheaply if you’re wrong, do not need a green light. Just make them. Reserve the asking for the genuinely expensive, hard-to-reverse calls. Most of life is reversible, and treating it that way is most of agency.
Ship before you feel ready
You will never feel ready, because “ready” is a feeling that arrives after the action, not before it. Set the bar at good-enough-to-get-real-feedback and release it there. A rough thing in front of a real person teaches you more in a day than a perfect thing in your head teaches you in a month. Perfectionism is often just permission-seeking wearing a nicer outfit.
Own the result out loud
When it works, put your name on it. When it breaks, put your name on that too, and on the fix. Ownership practiced in public is what converts scattered attempts into a reputation, and a reputation is what earns you the room to attempt bigger things. This is also the honest version: you can’t claim the wins without standing under the losses.
Keep a record of what your attempts taught you
After each rep, spend two minutes on one question: what did that teach me that I didn’t know before? This is the step that separates learned agency from the reckless kind. The record turns raw action into calibrated judgment, so that over time you’re not just braver, you’re righter. Bravery without learning plateaus. Bravery with learning compounds.
Do this for a season and something shifts that’s hard to describe until you feel it. The default setting of your life changes from “wait and see” to “try and learn.” The category of things you’ll simply do on your own expands. And because it’s the meta-skill, everything else you know starts getting used instead of sitting idle. The languages you learned, the tools you can wield, the taste you’ve built: all of it finally leaves the shelf, because agency is what takes it down.
The person still waiting to be told what to work on isn’t less talented. They just never practiced the one skill that turns talent into a life. You can start practicing it in the next hour. That’s the whole point. Nobody has to hand you the first rep.
