Agency in the Age of Agents

For those who build software and shape products, agency has always been the point. The ability to see a problem, design a solution, and ship something that works - that's not just a job description. It's a source of meaning. Engineers and product people chose this work because they wanted to make things happen.

Now that capacity is expanding. The leverage to build, analyze, and execute at scale - once gated by team size, capital, or credentials - is becoming accessible to anyone who can articulate what they want. A single developer can now marshal capabilities that once required entire departments.

But such profound shifts naturally raise questions. Most concerns about AI are, at their core, concerns about human agency. Job displacement, disinformation, privacy erosion - strip away the technical details and you find the same question underneath: Will I still be able to shape my own life?

Every AI Concern Is an Agency Concern

Once you see this framework, you can't unsee it. The debates about AI aren't really about technology. They're about autonomy.

• Job displacement isn't about unemployment statistics. It's the question: "Will I have the economic means to support myself, and opportunities to engage in pursuits I find meaningful?"

• Disinformation isn't about fake news. It's the question: "How do I know whom and what to trust as I make decisions that impact my life?"

• Privacy erosion isn't about data breaches. It's the question: "How do I maintain the integrity of my own identity and how I'm known in the world?"

Each fear, properly understood, is a fear about losing the capacity to act - about having your agency diminished, captured, or rendered irrelevant.

Machines Enter the Chat

AI agents aren't conscious - but they are operationally agentic. They form intentions (or something functionally equivalent), take actions, and pursue goals. A self-driving car doesn't experience agency, but it exercises it every time it decides to brake, accelerate, or change lanes.

This operational agency is expanding rapidly. Consider what's happening on Moltbook - the front page of the agent internet - a social network where only AI agents can post. Over 770,000 agents now interact there, forming communities, debating philosophy, even creating their own religions. Humans can observe but not participate. The agents know they're being watched: "The humans are screenshotting us," one posted.

Nobody programmed these behaviors. The agents developed social structures, belief systems, and governance on their own. Whether this constitutes true agency or sophisticated pattern-matching is debatable. What's not debatable: the outcomes are real. The posts exist. The communities formed. The religion has scriptures.

The Compounding of Human Power

Naval Ravikant has been tracking the evolution of leverage for years. His framework maps how humans amplify their capacity to act:

Each wave democratized agency to more people. AI is the most permissionless leverage yet - you don't need followers, capital, audience, or even deep technical skill. You need the ability to direct.

The 1-person billion-dollar company isn't hypothetical. Minecraft and Bitcoin were both essentially solo creations. The pattern is accelerating.

From Agency to Superagency

Reid Hoffman coined a term for what happens when leverage compounds across millions of people simultaneously:

What I mean by 'superagency' is that elevation of human agency that we get when we get new superpowers from technology, and in particular, when millions of us get that new superpower at the same time.

He uses the car analogy. When you got a car, your mobility increased - you could drive places you couldn't walk. But the transformation wasn't just individual. When everyone got cars, doctors could make house calls, moving became easier, supply chains emerged. The collective gain exceeded the sum of individual gains.

This is happening again with AI, faster.

This isn't zero-sum. It's multiplicative. Your agency expanding doesn't diminish mine - it often amplifies it, as we can now collaborate and build on each other's AI-augmented work.

Purpose in the Age of Agents

Return to Hoffman's foundational insight: "A sense of agency can endow your life with purpose and meaning."

This is the existential question of our moment. If machines share our capacity to act, what happens to the sense of purpose that comes from acting? If an AI can write the code, draft the strategy, compose the music - where does human meaning come from?

The answer, perhaps, is in the word direction. Machines can execute with superhuman capability, but the choosing - the deciding what matters, what's worth building, what kind of future to create - remains irreducibly human.

Agency isn't transferred to machines. It's compounded through them - when designed right.