The Anxiety of Tireless Workers

I got more done today than I could have accomplished in a week five years ago. And yet, I’m sitting here feeling like I didn’t do enough.

Welcome to the anxiety of tireless workers.

The Paradox

Here’s the thing about coding agents: they don’t get tired. They don’t need coffee breaks. They don’t context-switch poorly after lunch. You throw a task at them and — seconds later — it’s done. PR ready. Tests passing. Move on.

So you move on. And then you move on again. And again. And by the end of the day, you’ve shipped what would have been a solid sprint in the before-times, but your brain is screaming: “That’s it? That’s all you did?”

The bottleneck shifted. It used to be execution. Now it’s you — your ability to think, prioritize, decide, and feed the machine. And that shift is psychologically brutal.

The New Stress

This isn’t the old stress. The old stress was “I have too much to build and not enough time.” That stress had a shape. You could see the backlog, estimate the effort, and make peace with the constraints.

The new stress has no ceiling. When your agent can execute anything in seconds, the only limit is how fast you can think of things to execute. Every idle moment becomes a wasted opportunity. Every break feels like you’re leaving money — or momentum — on the table.

It’s the stress of abundance. The anxiety of having a factory that never stops, where the only assembly line waiting is the one inside your head.

The Irony

This very blog post exists because of that dynamic. I asked my AI agent to help me write it. She had it drafted in under a minute. And even that — the act of producing a thoughtful reflection on productivity anxiety — somehow feels like it wasn’t enough.

The irony is thick: I’m using the thing that causes the anxiety to articulate the anxiety, and I’m still not satisfied.

What I Think Is Happening

I think we’re experiencing a new form of cognitive dissonance. Our internal model of “a day’s work” was calibrated over years — decades — of human-speed execution. We have deep intuitions about how long things should take, how much effort should feel like enough.

When an agent compresses a day’s work into an hour, our calibration breaks. The effort didn’t feel proportional to the output. And because we’re wired to measure our worth partly through effort, the math doesn’t add up emotionally.

You know you accomplished a lot. But it doesn’t feel like it.

The Comparison Trap

There’s another layer: you start comparing yourself not to your past self, but to the theoretical maximum. “If the agent can do X in 10 seconds, and there are 28,800 seconds in a workday, then I should be able to…”

Stop. That math is a trap. You’re not a dispatcher optimizing throughput on a factory floor. You’re a human who needs to think, rest, context-switch, eat, and occasionally stare at the wall while your subconscious processes a hard problem.

But try telling that to the part of your brain that just watched an agent refactor an entire module while you were making mate.

What I’m Learning

I don’t have answers yet. But I’m starting to notice a few things:

1. Redefine “a good day.” It’s no longer about lines of code or tickets closed. It’s about decisions made, directions set, problems understood. The agent handles execution. Your job is judgment.

2. Rest is not waste. If the bottleneck is your brain, then resting your brain is literally maintaining the critical path. This is not laziness. This is capacity planning.

3. The feeling is wrong. Not morally wrong — just inaccurate. Like jet lag for productivity. Your body (mind) hasn’t caught up to the new timezone. Give it time.

4. You’re not competing with the agent. You’re collaborating. Comparing your output to the agent’s is like a conductor comparing their note output to the orchestra’s. Different job.

The Bigger Picture

I suspect this anxiety is going to be one of the defining psychological challenges of the next few years. As agents get better, faster, and more autonomous, the humans directing them will need to fundamentally rewire their relationship with productivity, effort, and self-worth.

We spent centuries optimizing for doing more. Now we need to optimize for thinking better — and being okay with the fact that thinking doesn’t look like working.

Today I didn’t write as much code as I wanted. But I made architectural decisions, reviewed agent output, caught edge cases a machine would miss, and set direction for tomorrow. And now I’m writing about it, which might help someone else navigating the same weird feeling.

That’s a good day. Even if it doesn’t feel like one yet.