Do You Have Self-Control? Agency in the Age of AI
Why agency matters, and how to build it
Draft: this essay is a work in progress, and I’m still revising it.
Lately, it seems like one idea just keeps reappearing in my life: the concept of agency. I’ve begun to notice it in every person I admire: a film student, one of 30 admitted to UCLA’s program, who really chases his dreams; a researcher pursuing the edge of what’s known; a young man in his 20s starting an AI company in the Bay Area. These people all had something in common. They knew what they wanted and actually took the action to make it happen.
The word for this is agency. I’d define it as healthy risk assessment plus consistent, goal-oriented behavior, i.e., knowing what you want, holding it in mind, working toward it even when progress is slow, and taking real risks to move forward.
It’s also, suddenly, the word for software. “AI agents” are systems that are given a goal, break it into steps, keep it in working memory, act, check their progress, and try again in an almost autonomous cycle. Reading that description, I realized it’s almost exactly the list of things I struggle to do myself, especially staying focused and following through.
That’s the strange part of this moment. We’re building machines with agency at the same time that more and more of us feel like bystanders: consuming other people’s work, admiring their wins, and staying inside our comfort zones. I’ve done this. I’ve tried to be close to agentic people, hoping some of it would rub off. It doesn’t. Knowing someone who acts is not the same as acting.
As AI (both chatbots and agents) competes for jobs, creative work, and even attention, the pressure to “maximize your potential” is everywhere. But that raises a real question: Is agency worth pursuing, or is it just part of the rat race? And, more importantly, how can we all maximize our agency in this newly realized competition against automated minds?
In this essay, I argue that agency is one of the most important traits to have in the modern day, regardless of where your starting point is. If you can turn an idea into reality, you have a strong edge over those who sit around overthinking and not acting. Some argue that adapting alongside AI, and relying heavily on AI tools for everyday life and work, is just allowing yourself to become part of the capitalist rat race. For me, it’s less about being against living a slow life and more about knowing what you want, taking into account all of the associated variables, running a simulation in your mind to see if those variables interact the way you want, and then choosing a path and acting on it. I don’t believe in free-floating and “just figuring it out” for too long.
Okay, so then, the next question is: how can one develop more agency? There are five main aspects of agency: one, knowing what you want; two, researching how to get closer to that goal; three, taking the first step, even with fear; four, making adjustments when necessary; and five, staying consistent with this loop of acting and revising until you complete the goal.
One: Direction, or Knowing What You Want
This is something most people my age struggle with because we have yet to figure out who we are, and therefore what would fulfill our goals, values, and emotional needs. So it’s important to first attempt to figure those things out, for example with a values chart: list what matters most to you, rank it, and notice which goals actually line up with it.
After making a values chart, start running down one direction, even with the lights out, and just see where it takes you. It’s okay to pivot later in your twenties. A wrong direction still teaches you more than standing still.
Two: Figuring Out How to Get There
Ask those around you. Don’t be afraid to network. Google it. Ask AI. Find someone who is two steps ahead of you and ask what they did. This is actually the one I struggle with the most, because asking for help can feel like admitting I don’t know what I’m doing. But nobody does at first. Research is just the cheapest form of action.
Three: Taking the First Step, Even with Fear
Personally, I like to think of attempting something like dating. If you fail, there are literally millions of other people out there who could be a better fit. If the date doesn’t go well, it’s not the end of the world. It’s about the risk/benefit ratio. Most first steps are like this: the downside is a little embarrassment, and the upside is that you’re finally moving.
Four: Adjusting Along the Way
No plan survives contact with reality. After you act, check in: is this getting me closer to what I want? If not, change one thing at a time instead of starting over completely. Funnily enough, this is exactly what AI agents do: act, check progress, try again. The difference is that we tend to take a failed attempt personally. A machine just updates. I’m trying to learn to be a little more like the machine here.
Five: Staying Consistent
Agency isn’t one big brave moment. It’s the loop of acting and revising, repeated until the project is done. Motivation fades, so it’s better to build systems instead: small daily steps, real deadlines, and people who hold you accountable. For me, this means making the next step so small that starting feels easier than avoiding it.
So, How Do We Compete with AI?
AI can follow a goal, but it can’t choose one that means something to it. That part is still ours: deciding what matters, taking risks with real stakes, and caring about how things turn out. So maybe competing with AI doesn’t mean outworking it. Maybe it means using AI as a tool for the parts of agency we struggle with, like research, planning, and breaking big tasks into steps, while keeping the choosing for ourselves.
TBH, I don’t have a neat ending, because this is the thing I struggle with the most right now. I know the five steps. I can write them down. But knowing the loop isn’t the same as running it, just like knowing someone who acts isn’t the same as acting. So maybe the conclusion is simply this: I’m publishing this essay before it feels finished. That’s my first step.