The End of Logistics

Most of what wears people down is not the work. It is the logistics around the work: remembering, tracking, scheduling, chasing, filing, checking it again. I have handed nearly all of mine to an AI and kept the deciding and the doing. The surprise was not that I got more done. It was that I stopped feeling stressed, with more going on than ever. This essay is about why that happens, what it costs, where it already broke on me once, and who gets to have it.

A confession

I hate logistics. I always have. I was good at them anyway, the way some people are good at parallel parking and still dread it.

By logistics I mean everything around the work that is not the work. The reminder to send the thing. The tab where you track whether the thing was sent. The calendar invite about the thing. The folder the thing goes in, the version of the thing with the right file name, the second look to make sure the thing actually went live. None of it is hard. All of it has to be held somewhere, and for most of my life the place it was held was my head.

Right now I run a one-person company with more going on than I would have believed a year ago: a game studio, these essays, free templates, a clothing label in development, music, consulting. A few months ago that list would have flattened me. Today it does not, and the reason is simple. I do not run the logistics anymore. The AI I work with all day does.

A normal day looks like this. I say what I want in a sentence, often misspelled. The thing happens. The scheduling, tracking, filing, deploying, drafting, reminding, chasing and checking all happen somewhere I do not have to look. I make the decisions and I do the parts only I can do. I only know what I need to know, and it works great.

What I did not expect is how it feels. I am happier and more relaxed than I have ever been, in the middle of the busiest stretch of my life, and I feel no stress. That is the part I want to understand, because it is not what the productivity pitch promises. The pitch promises more output. What I got was calm.

What logistics actually is

There is a name for this stuff in offices. Asana, the work-management company, calls it "work about work": time spent coordinating work rather than doing the skilled part. Its 2023 survey of 9,615 knowledge workers in six countries found that work about work took up 58% of their day (Asana, 2023). That number comes from a company that sells software to fix the problem, and it is people estimating their own time, so treat it as a rough picture, not a measurement.

The measured numbers point the same way. Microsoft looked at how people actually use its email, chat and meeting software and found that during core hours the average worker gets interrupted by a meeting, an email or a notification about every two minutes, and receives 117 emails a day (Microsoft Work Trend Index, 2025). Microsoft sells software too. But that figure comes from usage data, not a poll. Back in 2012, McKinsey estimated that managers, professionals and salespeople spent about 28% of the workweek on email and another 19% tracking down information (McKinsey Global Institute, 2012). Different methods, different decades, same shape. A huge share of the working day is spent keeping work organized.

It does not stop at the office door. The sociologist Allison Daminger interviewed both partners in 35 couples and found that running a household includes a whole layer of what she called cognitive labor: anticipating what will be needed, identifying the options, deciding, and then monitoring whether it got done (Daminger, American Sociological Review, 2019). She found that this work is often invisible, even to the person doing it, and that it is a common source of fights. People call it the mental load. It is logistics with nobody's name on it.

The writer Craig Lambert has a name for another slice: shadow work, the unpaid jobs we do on behalf of businesses and organizations, many of which someone used to be paid to do. Bagging your own groceries. Checking yourself into a hotel at a kiosk. Doing your own taxes (Lambert, Shadow Work, 2015). And the political scientists Pamela Herd and Donald Moynihan have spent years documenting administrative burden, the costs people pay just to deal with government: learning what applies to them, complying with the forms, and the stress of the whole process (Moynihan, Herd and Harvey, 2015). Their argument is that those costs are not neutral. They decide who actually gets help.

Put it together and you get a picture of modern life where a large part of everyone's attention goes to holding the work, not doing it. That is the part I handed off.

Why it makes you calm, not just fast

Here is my theory. Stress does not track the number of things happening. It tracks the number of things you are holding.

There is an old idea in psychology called the Zeigarnik effect: unfinished tasks stick in your mind better than finished ones. It is a famous result, and I want to be careful with it, because it is weaker than its fame. A 2025 meta-analysis by Romain Ghibellini and Beat Meier found no general memory advantage for unfinished tasks. What did hold up was the related finding that people tend to go back and finish what was interrupted (Ghibellini and Meier, 2025). So the unfinished thing does not necessarily get remembered better. It just keeps pulling at you.

The more interesting study is one from 2011 by E. J. Masicampo and Roy Baumeister, titled "Consider it done!" They gave people unfinished goals and found the goals leaked into everything else. People had intrusive thoughts about them while reading something unrelated, and did worse on an unrelated word puzzle. Then they let some people make a specific plan for the unfinished goal, and the leaking stopped. The goal was not done. It just had a plan, and the mind let go of it (Masicampo and Baumeister, 2011). This is one set of studies from one lab, so I hold it loosely. But it matches what I feel almost exactly. The relief does not come when the thing is done. It comes when I trust that the thing is handled.

Then there is attention residue. The management researcher Sophie Leroy showed that when people switch from an unfinished task to a new one, part of their attention stays behind on the old one, and they do worse at the new one (Leroy, 2009). Every open loop you carry is a little bit of you that is not in the room.

You may have heard a fourth explanation, decision fatigue: the idea that willpower is a tank and every small choice drains it. I am leaving that one out of my argument. The research behind it, called ego depletion, has not held up well. A replication across 23 labs and over 2,000 people found an effect close to zero (Hagger and colleagues, 2016), and a later test across 36 labs found no evidence for it in its planned analysis (Vohs and colleagues, 2021). It is a good story. It is not a solid fact.

So here is what I think is going on, stated as plainly as I can. When I held my own logistics, every project was a set of open loops, each one asking for a little attention, all day, whether I was working on it or not. Now the loops are held somewhere else, and I trust that they are. The number of things happening went up. The number of things I am holding went down. And the calm followed the second number, not the first.

Where I am wrong

The habit in this series is to say where I am wrong before where I am right. There is plenty here.

The better it works, the worse I get at noticing when it does not. In 1983 the engineer Lisanne Bainbridge wrote a short paper called "Ironies of Automation." One of her ironies is that automation leaves the human the tasks the designer could not figure out how to automate, including watching for the rare failure, which is exactly the job people are worst at. She wrote that even a highly motivated person cannot keep watching effectively, for something that rarely happens, for more than about half an hour (Bainbridge, 1983). Ten years later, Raja Parasuraman and colleagues showed it in the lab: when people had other tasks to juggle and an automated system was reliable all the time, they got worse at catching its failures than when its reliability varied (Parasuraman, Molloy and Singh, 1993). Later work found that this complacency shows up in experts too, and that simple practice does not fix it (Parasuraman and Manzey, 2010).

I have already lived this. Recently, one of my projects, a website and the production prep for a clothing line, sat untouched for four days. I assumed it was in progress. The AI had stopped and had not told me. I found out by asking. "Only knowing what I need to know" failed exactly once, and it failed quietly, which is the only way it could fail. As far as I can tell, all it cost was four days. But the lesson is the one Bainbridge wrote down forty years ago. If you hand off the holding, you have to keep some way of checking that it is still being held, and you cannot rely on feeling fine as the check. Feeling fine is the whole point. It is also the blind spot.

Responsibility does not move. There are still things only I can do, and some of them are set up that way on purpose. The AI could draft an email to a manufacturer I had never written to, but a safety layer would not let it press Send. Money, signatures and first contact with someone new stay with me. I think that is right. The end of logistics is not the end of accountability. When something goes out under my name, it is mine, no matter who tracked it.

I might be losing a skill. Psychologists call it cognitive offloading when you put something in the world, like a calendar or a phone, instead of in your head. Evan Risko and Sam Gilbert, reviewing the research in 2016, describe the trade honestly: it gets you past the limits of memory, and it can also mean you remember less without the tool (Risko and Gilbert, reported by ScienceDaily, 2016). Bainbridge said the same about operators: skills that are not used deteriorate. I was good at logistics. I am probably getting worse at them. I am fine with that trade for now, but I should admit that if the tool went away tomorrow, I would not be as good as I was.

Not everyone gets this. Today it takes money for the tools, some skill at delegating, and enough trust to let go. Those are not evenly spread. In a Pew survey published in early 2025, 55% of American workers said they rarely or never used AI chatbots at work, and the people most likely to use them were younger or had postgraduate degrees (Pew Research Center, 2025). That will change, but I should not write as if my week is everyone's week. And my guess, which I cannot prove, is that the people carrying the heaviest logistics, the ones filling out the forms Herd and Moynihan study, will be among the last to get the tools that would lift them.

What ends and what does not

So what actually ends?

Logistics ends. The remembering, tracking, scheduling, chasing, filing and formatting are exactly the kind of work a machine can hold, because they are rules and lists and follow-ups, not taste. My guess is that for a lot of people, a good share of the tiredness is this.

Judgment does not end. Deciding what to make, who to make it for, what is good enough and what is not, whether to say yes: none of that got handed off, and I do not think it should. If anything, it got more of me, because it is no longer competing with forty open loops for attention.

And I get some of my attention back. I spend it on music, on my wife, on writing these, and on building things for people who need them. That is not a productivity gain. You cannot put it on a chart. It is just a different life.

If you want to know whether this would help you, here is the test I would suggest. Take a piece of paper and write down everything you are holding in your head right now that is not the work itself. The appointment you have to book. The form you have to find. The person you have to remind. The thing you are pretty sure went out but have not checked. Do not sort it. Just write until you run out.

That list is what can end. Not all of it today, and not for everyone yet. But it is real, and it is heavier than it looks, and you have been carrying it the whole time.

Sources and a note on method

Every figure is linked where it appears, to the primary paper, the publisher's or author's own page, or the most direct report I could open. A few rest on reports rather than the full text: the Masicampo and Baumeister findings are from the article's published abstract; the Leroy, Hagger and Vohs findings are from their abstracts; Lambert's definition and examples are from his Publishers Weekly interview; Risko and Gilbert's review is described from ScienceDaily's report with their own quotes. The Asana and Microsoft figures come from companies that sell workplace software, and I have said so where they appear; Asana's is a self-report survey of 9,615 knowledge workers run in November 2022, and Microsoft's interruption and email numbers come from its own usage data. Herd and Moynihan's three costs are from their 2015 journal article with Carolyn Harvey; their book Administrative Burden: Policymaking by Other Means (Russell Sage Foundation, 2018) develops the same idea.

Left out on purpose: decision fatigue as an explanation, for the reasons given above; any claim that the Zeigarnik effect means unfinished tasks are remembered better, which the 2025 meta-analysis does not support; and any figure for how much time AI saves, which depends too much on who is asking and what they are doing to state honestly in one number.

The four-day stall and the email the AI could not send are from my own recent work, told as they happened. The research log behind this essay, with every source and the claims cut, is available on request. The earlier essays in this series are The Ego Death of Society, The Utility of Not Caring, Failing Confidently, Impatiently Waiting, Personal Bubble Universe, Why a Smart Machine Would Keep Us and Every Road Ends at the Same Sea.

I used AI to research and draft this, under my direction, with a standing instruction to tell me where I was wrong before telling me where I was right, and every claim was checked against its source before it went in. Fittingly, the AI also handled the logistics of publishing it.

Josh Solomon founded Known Quantity LLC in October 2026: an AI consulting practice for finance and revenue teams, the Gimble Gaming studio, three clothing labels (two with his wife, one his own men's line, Cordwell), and this essay desk. He uses AI to research and draft, under his direction, with every claim checked against its source. LinkedIn · Invite him to speak

Questions this essay answers