Some mornings, I feel like a program. My hand finds my phone before I'm fully awake, and the code just runs: unlock, scroll, refresh, repeat. It’s an input/output loop, a familiar script I execute without conscious thought. I’m not really making decisions; I’m processing a routine.

On other mornings, the pattern breaks. I notice the impulse to grab the phone and simply don't. Instead, I go for a run, make coffee, and watch the day begin perfectly. In these moments, I feel less like a script and more like an author, aware and operating with a sense of purpose.

In that first scenario, I’m more like a calculator — a tool, a program being run.
Yes, your calculator is a program.
Siri is more like an agent, still someone’s agent (yet capable of a goal).

But what about us? This isn’t just a computer science question; it's a deeply human one.

The Recipe vs. The Chef

To see the distinction clearly, think of it in the kitchen.

A program is like a recipe or recipe follower. It's a set of predefined instructions. Give it the ingredients (input), and it produces a dish (output). It doesn't improvise when you run out of an ingredient or adjust for a finicky oven. It just follows the steps.

An agent, however, is like the chef. The chef uses the recipe but has a broader goal: make a great meal. They adapt, make choices, and act autonomously to achieve the best outcome. They have agency.

A program is a tool that follows instructions.
Even better: a program is an agent without agency.

An agent is an entity that pursues goals.
In other words: an agent is a program with its own agency.

The space between following and pursuing is where the conversation gets interesting.

Agency and Autonomy (Where It Gets Messy)

When we talk about agency, we mean the capacity to make choices toward a goal.
When we talk about autonomy, we mean the freedom to act without constant oversight.

But here’s the uncomfortable question: do humans always have agency?

Psychologists argue about this constantly. Are our “choices” really free, or are they the outputs of biology, upbringing, and environment? Philosophers add fuel to the fire: if an AI follows rules but adapts to new situations, is its “agency” any less real than ours? Or are we just more comfortable calling our patterns free will and their patterns programming?

If free will is just a convincing illusion, then maybe the only real difference between us and machines is the complexity of our code.

The Human Autopilot

This lens exposes how often we slide into script. Some scripts make us efficient by saving us so much time and decision fatigue. Some just pull us into dreadful aimlessness and endless pits. Let's some of our autopilot scripts:

• The mindless commute to work, driven by habit.
• The conversational loops (“How are you?” “Fine, you?”).
• The reactive emotions that flare up in traffic.
• Worst of all: the endless doomscrolling.

When an AI adapts its strategy to win a game, is its agency fundamentally different from our own when we act on ingrained habits? The line blurs, and that’s exactly why this debate matters.

A Spectrum of Agency

“Program” and “agent” aren’t binary switches; they exist on a spectrum of complexity. Computer science offers a simple map:

• The Thermostat (Simple Reflex): Reacts to a single data point (temperature) without memory or planning. It’s our knee-jerk reaction.
• The Roomba (Model-Based): Builds a mental map of a room to remember where it's been. It’s our ability to use context before we act.
• The GPS (Goal-Based): Doesn’t just react; it plans a route to get you from A to B. It’s our conscious decision to work toward an objective.
• The Recommendation Engine (Learning): Observes, learns your preferences, and adapts to better serve its goal. It’s our capacity for growth.

Each step up this ladder introduces more flexibility and autonomy.

Final Thoughts: A Lens, Not a Label

Perhaps “program” and “agent” aren’t fixed identities, but modes we all shift between throughout the day — sometimes even moment to moment.

We all have our scripts — some helpful, some harmful. The deeply ingrained habits and reactions that save us energy and help us navigate a complex world… or slowly drag us into oblivion. But we also have moments of clarity and intention, where we step outside the script and make a deliberate choice.

The machines we build are the same. Many “AI agents” are just elaborate programs, while some simple algorithms learn and adapt in ways that feel genuinely agent-like.

The point isn’t to judge one mode as better than the other. It’s to recognize that both exist within us.

The interesting question, then, isn’t if you are a program or an agent. It’s…

When are you each?
When are you able to be aware?