## A Moment from Everyday Life
The other day I was standing in the kitchen. Dinner. Meanwhile an agent is running on the laptop, “just quickly” researching, structuring, drafting. I put away plates, do a couple of tasks, glance at my phone. My attention is on the open question: is the agent done? Is there already something there that I should approve, correct, or steer further?

And I noticed something about myself that I hadn’t known before, or at least hadn’t felt this clearly: this kind of “open loop” doesn’t just sit in the back of your mind as a to‑do. It’s active. It’s running. And it can jump back at any time and demand a decision.

I’ve started calling it: **Open Loop Life**. And with agents: an **Agentic Open Loop Life**.

This is an observation from a place of enthusiasm: I believe this is one of the most productive technologies we’ve received in decades, and its everyday usefulness is exactly what makes the problem relevant.

## What is “Open Loop Life” anyway?
A large part of daily life consists of open loops: things that were started but not finished. An open item in a project. An unanswered message. A thought you didn’t get to the end of. An “I should …”.

These loops have one property: they stay active even when you’re not actively working on them. They pull attention away. Not necessarily loud, often quiet. Like background noise.

And we all know the countermeasure that’s been preached for years because it actually helps: externalize. To‑do lists. Reminders. Kanban boards. “Getting Things Done.” Lists relieve the brain because they hold what you want to let go.

That works surprisingly well, up to a point.

## What changes with AI agents: from static to dynamic open loops
With a classic open loop, the pattern is relatively stable: “I still need to do X.” You can put it on a list. You can park it. The thing stays open, but it’s passive.

Agentic systems turn that into something else: “X is being worked on somewhere and can tip into an interaction moment at any time.”

That sounds like a detail, but psychologically it’s a different category. You are thinking about the task and the unknown moment when it returns. This creates a live relationship with a process that has its own tempo.

And exactly that “not on my tempo” is the core. The agent finishes when it finishes. Maybe in two minutes. Maybe in ten. Maybe at a moment when you’re in the middle of everyday life.

That’s the first break: temporal unpredictability.

The second break is at least as relevant: the difference from doomscrolling.

## The Difference from Doomscrolling
By now we’re good at naming doomscrolling as a problem. Passive loops: open the feed, keep scrolling, quick dopamine, little substance. You rarely feel better afterward.

Agentic open loops look similar from the outside. In content, something else is happening: you check because something might be there that has real value.

A summary that saves you two hours. A draft that sharpens a concept. A list of options you would have had to research laboriously. A structure that suddenly makes a diffuse topic tangible.

And that’s where it gets tricky: the checking doesn’t feel like distraction. It feels like responsibility. Like leverage. Like “I should quickly look because this is important.”

That makes these loops so powerful, and so hard to regulate.

## Why agents create such a strong pull
A few mechanisms come together:

* **Variable completion:** Output arrives at irregular intervals. That almost automatically creates “progress tracking” in your head, because you want to catch the moment when it becomes interactive again.
* **High potential benefit:** The next check could deliver a lot of value. With a feed, the benefit per check is usually small. With agents it can be large.
* **Contingencies and limits:** If you have a budget (time windows, calls, tokens, quotas), a second layer quickly appears: “Am I using this optimally? Am I leaving output on the table?” This meta‑loop is an amplifier.
* **Low start costs:** A prompt is easier than a project plan. That means you start more loops than you could ever manage cleanly. It feels like momentum at first and like backlog later.

Put together, something paradoxical emerges: the system makes “quickly checking” feel plausible and creates restlessness alongside higher productivity.

## Decision Debt: the hidden price of new speed
Agents can generate material incredibly fast. Variants. Alternatives. Pros and cons. Drafts. Steps. All in minutes.

But in the end a human sits there and has to evaluate, prioritize, decide, and integrate. That isn’t fully delegable.

That creates **Decision Debt**: a growing debt of decisions you still have to make because so much output was produced so quickly.

This is a time problem and an attention problem at once. You can run multiple agents in parallel, but you can’t meaningfully process results in parallel. Attention is serial. Life is serial. And that’s where friction emerges.

## Orchestration Tax: when coordination becomes the real work
In the classic work world the ratio was often clear: you do the work, you deliver the result.

With agents, the ratio shifts: you orchestrate. You steer. You set tasks, correct course, define “done,” check quality, add context.

That is valuable work, but it’s a different kind of work. And it has a tax: the **Orchestration Tax**.

This tax isn’t dramatic. It’s frequent. It shows up as constant switching between the life world and the agent world. Rarely large, constantly recurring. And “over and over” is often the real problem.

## Human Event Loop: the human as scheduler
Technically, it’s almost absurdly fitting: you are the **Human Event Loop**.

An agent fires events: “Here’s an intermediate result.” “I need a decision.” “Should I do A or B?” “I made an assumption; does it fit?”
The response can wait, yet the possibility remains palpable.

That’s the feeling I noticed so clearly in everyday life: even when I’m doing housework or stuck between two professional topics, a part of me stays in standby.

The standby mode appears because availability suddenly feels objectively justified.

**Why this is relevant right now**
I’ve long had that typical nerd thing: small ideas in my head that I’d like to program “sometime.” A small app. A tool. A little experiment that seems harmless but realistically eats time: architecture, code, tests, UI, details, iteration.

If you have a job and a family, your head eventually says: this isn’t necessarily hard. It’s mainly big. The entry cost is so much uninterrupted time that it hardly finds a natural place in everyday life. And then something many people know happens: you let it sit. Weeks. Months. Years.

With agents, that’s tipping right now.

The entry point shifts to the question: “What’s the next sensible step an agent can prepare while I live my life?”

That’s a massive reversal. And at its core it’s positive. It lowers hurdles. It turns “too big” back into “doable.”

But exactly because of that, a new pattern emerges: when start costs drop, the number of started loops rises. It becomes easier to kick off ten things you would never have begun before. It’s fascinating. And it brings us to a new boundary that we as humans still have to explore.

## The central difference from earlier open loops
In the past, the bottleneck was often implementation capacity.

Today the bottleneck is increasingly attention, decision, and integration.

In the past, open loops stressed you because they meant work.
Today open loops can stress you even though the work is “already running.” Maybe precisely because of that.

A sentence that sums it up well for me: when work becomes parallel but attention remains serial, new tensions arise.

## Why This Goes Beyond Productivity
Because everyday life is involved too.

About the small windows in which you could actually recover. About the walk that could happen “in a minute,” but then three small follow‑ups arrive, a result, another loop, another prompt, and suddenly the time window is gone.

And the tricky part is: it doesn’t feel like waste. It feels like progress.

Doomscrolling destroys time by emptying it.
Agentic open loops can destroy time by fragmenting it.

The result can be similar: less rest, less clarity, fewer real pauses. With the difference that you feel like you’re doing something meaningful while it happens.

## Terms that help me grasp the phenomenon

* **Open‑Loop Inflation:** because starting is cheap, you open more loops than you can reliably close.
* **Decision Debt:** output is fast; decisions are the bottleneck.
* **Orchestration Tax:** coordination, review, and steering become the primary load.
* **Quota Anxiety:** limits and quotas create an additional optimization loop in your head.
* **Human Event Loop:** you feel like the constantly available interface between running processes and real life.

## What I want from this text
I am not trying to sell a solution here. I even believe there won’t be one single solution. New habits, new norms, and new forms of digital self‑leadership will have to emerge that fit agentic systems.

The first step is naming the problem before it shapes us unnoticed.

Because when something works very well, it quickly becomes invisible. It becomes normal. You get used to living in half‑open loops all the time. You take that as the price of progress. And maybe only notice late that you hardly have any truly closed loops in your day, those quiet moments in which nothing could “come back any second.”

Maybe that’s the core question of the next years: how do we shape a life with open to‑dos and running processes?

And one more question that captures the contradiction well for me: how do we use the leverage of these systems without becoming the constantly available interface ourselves?