Robin Schooling

Who’s Flying the Plane?

Those of us of a certain age once kept a small directory in our heads, stocked with the phone numbers of our best friend, our grandparents, Domino’s, and the cute guy/gal we crank-called during slumber parties. And then, when our phones began remembering the numbers for us, that directory was cleared out so gradually that most of us can’t say when it happened or what moved into the vacated space in our brains.

That trade-off is relatively harmless; the contacts app on my iPhone rarely fails when it’s time for me to call the dogs’ groomer. But the same sort of switcheroo looks considerably different when the skill being traded away is the ability to fly an airplane, which is precisely what worried an American Airlines captain named Warren Vanderburgh nearly thirty years ago.

Following the Line

Vanderburgh, who spent 27 years in the Air Force before flying for American, was asked by the airline in the mid-1990s to figure out why so many accidents, incidents, and violations seemed to trace back to what came to be called “automation dependency.” In 1997 he stood in front of a class at the American Airlines Training Academy in Dallas and delivered a presentation he titled Children of the Magenta Line, borrowing the name from the magenta course line that a plane’s flight computer draws across the cockpit display. His concern was that pilots had grown so accustomed to following that line – managing the automation rather than flying the aircraft – that their own hand-flying skills were wearing away. And his team’s analysis attributed 68% of the problems they reviewed to automation mismanagement.

What makes the talk worth watching today (it lives on YouTube in all its late-nineties video glory) is that Vanderburgh was no Luddite. He laid out three levels of automation, from manually flying the plane, to the autopilot, to the full flight management computer, and argued that the professional skill lay in choosing the right level for the moment and being willing to step down a level when the situation called for it. The problem, as he saw it, was a culture that pushed pilots to always operate at the highest level of automation – a culture, he pointed out, that the industry had built for itself.

The Magenta Line Comes to the Office

The office version of the magenta line is the AI-generated first draft. It’s the meeting summary we skim in lieu of attending the meeting, and it’s the tidy synthesis of a forty-page report we never open.

And it arrives with a similar promise of reduced workload.

By their own account, employees have started to feel the drift Vanderburgh described. In a survey of 2,500 workers and IT leaders released this spring by GoTo and Workplace Intelligence, half of employees said they rely too much on AI, 39% said that reliance is eroding their skills and making them less intelligent, and 28% said they’ve begun trusting AI more than their own judgment. (That the research was commissioned by a company selling AI-enabled software gives the findings an almost confessional quality.) Six in ten also reported feeling pressured to use these tools in the name of productivity, which suggests that the culture pushing everyone toward the highest level of automation is not confined to cockpits.

Measured Decline

Self-reported worry is one thing and measured decline is another, and medicine has started producing the latter. In 2025, gastroenterologists Krzysztof Budzyń and Marcin Romańczyk of the Academy of Silesia in Poland and their colleagues published a study in The Lancet Gastroenterology & Hepatology examining what happened after four endoscopy centers introduced AI tools that flag polyps during colonoscopies. When those same doctors later performed colonoscopies without the AI, their adenoma detection rate fell from 28% to 22%. The researchers came across the finding by accident and were careful to frame it as a hypothesis about deskilling rather than a verdict, but a six-point drop among experienced physicians is not the sort of number we want to merely shrug off.

None of this would have surprised Lisanne Bainbridge, the cognitive psychologist whose 1983 paper Ironies of Automation remains one of the most cited works in human factors research. Her central irony is that automation absorbs the routine work and leaves humans responsible for the rare, difficult exceptions, which happen to be the moments when skills that have gone unused for months are least likely to show up.

The better a system performs day to day, in other words, the less prepared its human operators are for the day it doesn’t.

Who Learns to Hand-Fly?

For HR, the most uncomfortable part of the magenta line problem has less to do with the experienced workers sitting in our companies today than with the people coming up behind them.

Carl Hendrick, who spent eighteen years teaching in classrooms and now works on the science of how people learn, has been circling these ideas for a while – his March piece carries Bainbridge’s ironies straight into the modern knowledge-work office – and this summer he borrowed Vanderburgh’s phrase for a terrific essay in his Substack, The Learning Dispatch, about what chatbots are doing to students. In it, he draws a distinction that ought to make anyone responsible for workforce development a little uneasy. The pilots Vanderburgh worried about had learned to fly by hand long before the flight computers took over, so what they lost was ready access to knowledge they already possessed.

Novices who reach for the chatbot before building any foundation of their own, Hendrick argues, never acquire that knowledge at all. And nobody can drop down a level to fly a plane they were never taught to fly.

Skipping Ground School

Hendrick’s subject is the classroom, but the argument travels easily into the office. The Polish endoscopists, like the pilots, had years of unassisted procedures behind them before the AI arrived.

That 24-year-old claims analyst you’ve just hired, on the other hand, may never write the clunky first draft, reconcile the spreadsheet line by line, or struggle through the first set of investigation notes that once built judgment through sheer repetition.

We tend to think of AI deskilling as a loss – the gradual wearing away of something a person once had. But for a growing share of the workforce, the skills aren’t eroding at all … they can’t erode because they were never laid down in the first place. Medical educators have been wrestling with the same problem among residents and trainees, and the vocabulary has made its way into the New England Journal of Medicine: deskilling describes the loss of a skill someone already had, while “never-skilling” describes the failure to acquire it at all, because AI did the work during the period when the skill should have been forming (not to be confused with its unfortunate cousin, “mis-skilling,” in which the trainee learns the AI’s mistakes as if they were facts).

And our HR processes and systems aren’t built to notice these things either. Performance management rewards output, and the magenta line produces a great deal of output. Competency models assume that a skill, once demonstrated, stays put. And most L&D programs (and budgets) are planned around adding new capabilities rather than protecting existing ones from atrophy.

Manual Override

Aviation eventually responded by encouraging pilots to step down a level and hand-fly on purpose, and I think it’s fair to wonder what the workplace equivalent might look like. Perhaps a periodic exercise in drafting reports from scratch? Or a performance conversation that asks someone not only what they produced but whether they can explain how they produced it?

Pilots do have one advantage the rest of us lack though, because when an autopilot disconnects, an alarm sounds and the crew knows the plane is theirs to fly. When Claude or Grok or ChatGPT hands the work back to us – with a fabricated citation or a plausible but wrong summary – it doesn’t point out that it’s wrong. No alarms are buzzing on our desktops.

I’m not sure how many of us can recognize those moments when they come … let alone remember what to do with the controls.

“No White after Labor Day” and Other Authorless Rules

Nobody knows who made the rule. There’s no founding document showing us there was an etiquette board that convened in 1903 and took a vote. No one has found evidence of a memo that was circulated among the summer houses of Newport announcing that white linen would be retired on the first Monday in September. Historians have theories of course, and the most persuasive one has very little to do with fabric weight or autumn mud.

White was what you wore when you weren’t working.

You wore it when you had somewhere cooler to be for three months and no particular reason to think about what a day’s labor does to a hemline or the cuffs of your pants. This rule that ended the season became a way for the people already inside the room to sort out who shouldn’t be allowed to enter the room. You either followed the rule or you didn’t — and knowing who didn’t, or couldn’t, follow the rule was entirely the point.

The rule has now outlived its own explanation by roughly a century. Vogue has pronounced it dead more than once. Every September though, somebody’s grandma still says something about people who shouldn’t be wearing white shoes.

I think about this a lot at work. But not because of the shoes.

The Unwritten and the Unexplained

Every organization I’ve been part of has carried a set of rules whose origins were lost in the same way — not written down anywhere anyone can find, not defensible under even gentle questioning, and enforced with a gravity wildly out of proportion to anything at stake.

The expense threshold set at $25 because of an incident in 2011 concerning Rory from Facilities that nobody can describe but everybody vaguely remembers. The dress code prohibition on open-toed shoes at a corporate office where the primary hazard is the carpet. The standing rule against scheduling anything before 9:30 AM, which has now survived by five years the departure of the executive who wasn’t a morning person. The PowerPoint template with the mandated shade of blue and the very specific hex code that’s not even in the company logo.

Ask about any of them and you’ll get a pleasant answer with no defensiveness whatsoever (because nobody really feels the need to defend anything). After all… “It’s just how we do it here.” (shrug)

But… whatever these rules were originally built to do, what they do now is sort and separate people.

Who Followed and Who Didn’t

Learning these types of work rules is how you demonstrate you’ve been around long enough to have learned them, and following them without comment is how you signal that you’re one of us. Which is why the new hire who asks “why do we do this?” is being more disruptive than she realizes. She isn’t challenging the rule so much as pointing out that there’s no there there and nothing is holding it up. And no one wants to hear that.

Now HR is not an innocent bystander here; we’re frequently the ones who wrote the rule down and memorialized it. Somebody did something ill-advised and a policy got drafted in response. The policy then outlived the person, the circumstance, and eventually anyone’s memory of why it had seemed necessary. And now it continues to survive in the department SOPs or the employee handbook — seemingly forever.

The complication is that not every unexplained rule is a white shoe in November; some of them are substantive and necessary. The requirement that feels most arbitrary is occasionally the only thing standing between the organization and a regulator, or between the organization and something genuinely bad that happened once and has not happened since. The challenge is that from the outside, structural and decorative look identical. Both are inherited, both go unexplained, and telling them apart takes more effort than we often bargain for.

Because without that discernment, we keep these rules — often for the same reason the white shoes go back into the closet come Labor Day. Not because anyone can explain them, but because asking about them may cost more than complying.

*****

Back in My Day

I vividly recall the precise orange of aerosol cheese sprayed straight from a can. I remember the crinkle-then-release of a Ding Dong freed from its foil, the pizza bagels that lived in the freezer and tasted like a sleepover, and the tinny cheer of Saturday-morning cartoon theme music playing in the background. None of this serves any purpose beyond pleasure. I don’t especially want the spray cheese back — I’ve been told that processed dairy foods are bad for me — and I certainly don’t want to be eight again, subject to bedtimes and the tyranny of the adults.

What I do want is the memory itself, albeit held at a safe distance, since I’ve already had the pleasure (or misfortune?) of living through it.

This is, by most accounts, exactly how nostalgia is designed to work. Constantine Sedikides, the University of Southampton psychologist who has spent decades studying the phenomenon, has found that it tends to arrive precisely when we could use it — triggered by loneliness, or by a flat, unmoored present — and that it responds by connecting us back to a version of ourselves we still recognize. It restores what he calls self-continuity: the sense that the person eating a Ding Dong in front of Saturday cartoons and the person answering emails at fifty-six are, improbably, the same person. His research also links the feeling to a restored sense of purpose and greater generosity toward the people around us. Nostalgia, in other words, isn’t about the Ding Dongs and pizza bagels; it’s about the proof that one’s self has existed in a continuum both then and now.

There is, of course, also a workplace version of nostalgia, and it’s a genre all of us have heard and seen performed live. This performance is often given by someone three decades into their working life and one glass of wine into the company holiday party — “ya know Beth, back in my day…”

And we’re off.

The Sepia Toned Highlight Reel

Once upon a time, people were in the building. You could walk down the hall and get an answer. You learned the job by sitting next to someone who already knew it. The phone rang, you picked it up, and the conversation ended when one of you hung up. Meetings happened around a table. You knew who was having a rough week because you could see it.

Certainly, some of that is true. Proximity did make certain kinds of teaching easier and shared physical space created a kind of ambient knowledge that a chat thread will never replicate.

But something happens with this sort of memory that never happens when we reminisce about watching Scooby-Doo, Shaggy and the rest of the gang un-mask a villain. Childhood nostalgia and workplace nostalgia may appear similar on the surface, but there are important distinctions. Childhood nostalgia doesn’t ask us to change our present behavior; nobody is proposing that we install rotary phones back on our kitchen walls.

Workplace nostalgia, on the other hand, is rarely idle and undemanding. It tends to arrive, ready for an argument about what today’s employees are supposedly missing. It also brings with it some unspoken opinions about who ought to be in charge of restoring things back to their former (implied) glory. “Back in my day, people had a work ethic” is not a fond memory so much as a revisionist personnel policy from the days of Phyllis Schlafly.

What Lands on the Cutting Room Floor

There are memories we excise from the workplace highlight reel running in our mind, and we edit them the way we often edit our childhood memories: selectively, and in our favor.

Few people wax nostalgic about the ashtrays on the conference table or the dress code that banned women from wearing pants in the office. The person fondly recalling the open office door was very often the person behind the door, and the company where everyone “knew their job” was really only marvelous for whoever had assigned the roles in the first place. And the three-martini lunch sounds fun and madcap and exotic right up until you remember who wasn’t invited to tag along to the restaurant.

None of this means our fond memories and feelings are dishonest. Sedikides points out that nostalgia is not a synonym for delusion and people mostly know, on some level, that they’re looking at a highlight reel and not a minute-by-minute blow-by-blow record that’s being sworn to under oath. The trouble starts when reminiscing over the highlight reel moves us from saying “it felt simpler then” to “maybe we should just undo whatever has made it more complicated.”

What often gets lost if we move along that continuum is any evaluation of what will happen to the people standing just outside camera range on that highlight reel.

Off Camera

Much of what gets mourned as the old workplace is never really about the workplace at all. I think what we’re mourning is the version of ourselves who occupied that workplace of yesteryear — the version that was less tired, less responsible for aging parents, and much less aware of how many ways our jobs and careers can be upended.

Once upon a time that orange spray cheese tasted great because we were eight years old. And back in the day the workplace seemed simpler because we hadn’t yet learned how many ways things could get fucked up through no fault of our own. Which leaves the real question for me: can we ever replicate the feelings of awe and wonder of an eight-year-old who thought a can of spray cheese was magic?

Shifting Work: How Micro is Micro?

I have a working theory about workplace trends: assign any behavior that has persisted for decades a hyphenated descriptor, and within a year HR will be issuing formal guidance on it.

Let’s examine micro-shifting and the sudden authority with which employees now invoke it. Working parents, after all, have been negotiating lunch hours long enough to squeeze in a pediatrician appointment since long before anyone reading this was hired. What’s genuinely new isn’t the behavior – it’s the label. And a label is worth putting to the test.

Start with the obvious test: a minute spent answering a text is unambiguously micro. Fifteen minutes spent dropping something at the post office is defensibly micro. But a color, cut, and blowout runs closer to three hours, and I have yet to meet a single person who describes that as a quick break. A mani-pedi, with the requisite quarter hour for the polish to properly set, edges toward ninety minutes. Somewhere along the way, it seems, the word “micro” stopped describing duration and started describing level of permission. It’s an interruption that gets to call itself ordinary, regardless of how long it runs, provided it’s bookended by work on either side. Or so, perhaps, we assume.

Which raises some more interesting questions like what occupies the space between these work shifts? Does it matter? Provided the work still gets done?

The Full Menu

Once I began cataloguing it, the way one might if drafting an actual HR policy on the subject, the range of breaks required to keep a personal life functioning turned out to be considerably wider than the term implies. There’s going to the dentist. There’s taking Mr. Crumples to the vet whenever his ears resume doing that one thing again. There’s the grocery pickup timed to outrun the after-school scramble and the need to return my library book before I get a late charge. And then there are the larger blocks: the color and cut, the wedding-adjacent mani-pedi, the once-a-year ophthalmologist appointment that inevitably runs long. None of these qualify as a “break” in the traditional sense – the fifteen-minute variety once sanctioned by a poster in the breakroom or a clause in the employee handbook. Strung together, this resembles, instead, a second life – running concurrently with the first and held together by whatever bandwidth happens to remain.

None of this comes free, however, simply because any single instance of it is brief. The organizational behavior researcher Sophie Leroy has spent the better part of two decades documenting what she calls attention residue: the finding that when we move from one task to another, some portion of our cognitive attention remains lodged in the task we just left, regardless of how deliberately we try to leave it behind. The effect doesn’t appear to discriminate by scale, either; her research suggests it registers as readily during a brief pause for an unrelated errand as it does during a full change of major projects.

And that is a considerably more elegant way of saying what most of us already suspect from experience: that returning from ninety minutes at the hair salon and dropping straight back into a budget spreadsheet asks more of the brain than the calendar notation “out of office” accounts for. The true expense of micro-shifting is never really about stepping away or heading for a walk around the block or doing the school-run; it’s the tally of how many times a day we ask our own attention to fully disengage, and then, moments later, fully re-engage somewhere else.

Is There a Limit?

I think there is, though its precise shape is fuzzy. Organizations with defined core hours – say, everyone expected to be reachable between 10 and 3, with the outer bands of the day left open for whatever life requires – seem to manage this better than the ones still operating on the fiction that everyone is equally available from 9 to 5 and equally invisible outside it.

But while core hours help considerably with coordination, they don’t really solve for energy. What happens when your best two hours of the day – the hours in which your brain is genuinely equipped for the difficult, unglamorous thinking your job requires – land squarely inside the organization’s core hours? And then those same core hours happen to contain the only available therapy appointment, the only DMV slot without a two-hour line, and the only salon opening before the holidays? Core hours don’t resolve that particular collision; they simply relocate it.

Look – I’m generally in favor of micro-shifting. I love it, in fact. I’ve also had the good fortune to work for organizations, and in roles, where it was genuinely available; not everyone in every job has that luxury. Micro-shifting promises to let the work bend around a life rather than the reverse, and I remain in favor of that.

No matter what we end up calling it next.

—–

photo credit: Public Health Image Library from the Centers for Disease Control and Prevention (CDC)