Skip to content
All writing
Opinion · 2026 · 07 · 3 min

Working with AI agents doesn't remove the mental load, it transforms it

The real challenge of working with AI isn't the model of the week, it's how we manage our energy and time in a reality that never rests. How I'm organizing work in loops of intensity and rest.

GA
Gus Aragón
Founder, Oktsec
Key takeaways
  • The real challenge of working with AI isn't the model of the week, it's how you manage your energy and time in a reality that never rests.
  • Working with agents doesn't remove the mental load, it transforms it into another energy that is sometimes more dopamine than real building.
  • A scheme that's working for me: loops of weeks of intensity followed by a week of rest, to recover context and perspective.

Beyond whether the topic of the day is Fable, GPT-5.6, GLM 5.2, the agent tooling or the providers, for many of us who work in this the real challenge ahead is a different one: figuring out how to manage our energy and our time against a reality that asks no permission and gives no rest.

Maybe it's the most basic and human problem of all. Personally, it matters more and more to know when to rest, when to stop, and when to step back to recover context.

Working with agents doesn't remove the mental load. It transforms it into another kind of energy, more powerful, that at times has more to do with dopamine spikes than with what I'm actually building. You kick off a task, the agent runs, you jump to another, you review, you fix, you launch again. The loop is fast, and it's easy to mistake motion for progress.

Living this all semester from the open-source trenches is forcing me to change how I organize work. One scheme I'm exploring is working in loops: weeks of intensity followed by a week of rest. I put it into practice in mid-June and, so far, I'm happy with the result.

My contribution graph from February to June, with peaks of intensity and quieter weeks My contributions from February to June. You can see the peaks of intensity and the valleys of rest.

The rest week isn't only for recharging. It's for recovering the perspective you lose when you're inside the problem all the time. When you stop, it gets clearer what you were actually building and what was just motion.

I don't have a formula, and this will probably change. But for the first time in a while, I feel like I set the pace, not the novelty of the week.

Frequently asked
Does working with AI agents reduce mental load?
It doesn't reduce it, it changes its shape. The loop of launching tasks and reviewing results is so fast that it's easy to mistake motion for progress.
How do you organize work with agents to avoid burning out?
A simple scheme: loops of weeks of intensity followed by a week of rest, which is more about recovering perspective than recharging.
Keep reading