The longest, most boring, most dangerous phase of any long-haul flight is the part where nothing is happening. Six hours over the Atlantic, autopilot holding altitude, and the two people in the cockpit are expected to just... watch. Not react, not problem-solve — watch. And the industry figured out decades ago that asking a human to maintain peak alertness for six hours of nothing is not just unrealistic, it's actively dangerous. Daniel's been thinking about this, and he sent in a whole set of questions. He wants to know why aviation treats humans as part of the safety system rather than the weak link, why hypervigilance is actually a liability, what specific systems manage low-vigilance states during cruise, and — this is the part that matters for everyone who's not a commercial captain — what we can steal from this approach for managing our own awareness day to day.
The core insight that flips the whole conversation is that Human Factors isn't about eliminating human error. It's about designing for the human that actually exists — with all our cognitive limits, our fatigue curves, our tendency to zone out when nothing's stimulating us. Aviation accepted something most industries still won't admit: the operator is part of the system, and the system has to be built around how that operator actually functions, not how you wish they functioned.
Which is a strange thing for an industry to admit, when you think about it. "We put two people in charge of four hundred lives and several hundred million dollars of aircraft, and they're going to get bored and tired and make mistakes, so let's plan around that."
It sounds strange until you look at what happens when you don't plan around it. Human Factors as a formal discipline in aviation really crystallized after a series of disasters where the plane was fine and the humans broke down. James Reason — British psychologist — developed the Swiss Cheese Model in nineteen ninety. The idea is that you don't try to make any single layer perfect. You stack multiple layers of defense, each with its own holes, and the accident only happens when the holes in every layer line up. A pilot misses a scan. The co-pilot is distracted. The warning system is set to the wrong threshold. The controller is overloaded. None of those alone causes the crash. All of them together do.
So you're not trying to build a perfect pilot. You're trying to build a stack of imperfect things that catch each other.
And the model applies to the human as much as to the machine. The human is one of those layers, and you design around their known failure modes. Fatigue. Distraction. Complacency. The Yerkes-Dodson law — this goes back to nineteen oh eight, two psychologists studying how arousal affects performance in mice — shows that performance peaks at moderate arousal and drops off at both the low end and the high end. It's an inverted U. Too little stimulation and you're bored and miss things. Too much and you're anxious, tunnel-visioned, making errors.
Nineteen oh eight. We've known this for over a century and most workplaces still operate like the goal is maximum arousal at all times.
Most workplaces don't have a crash that kills five hundred and eighty-three people to focus the mind. Which brings us to Tenerife.
Nineteen seventy-seven.
March twenty-seventh, nineteen seventy-seven. Two Boeing seven-forty-sevens collided on the runway at Tenerife in the Canary Islands. Five hundred and eighty-three fatalities. Still the deadliest accident in aviation history. And the planes were fine. The weather was deteriorating but manageable. What killed those people was a cascade of human factors: fatigue, stress from a diversion, ambiguous language over the radio, and a captain whose authority was so absolute that his co-pilot — who had doubts about the takeoff clearance — didn't challenge him forcefully enough. The KLM captain started his takeoff roll without clearance because he thought he had it, and the Pan Am aircraft was still on the runway.
The co-pilot knew something was wrong.
The flight engineer actually asked, "Is he not clear, that Pan American?" And the captain said, "Oh, yes." And that was it. Three people in the cockpit, two of them uncertain, and the hierarchy flattened the uncertainty. Tenerife is the reason Crew Resource Management exists. NASA formalized it in nineteen seventy-nine. United Airlines adopted it in nineteen eighty-one. The core idea is that the cockpit is a team, not a hierarchy, and every crew member has both the authority and the obligation to speak up.
Which sounds obvious now, but at the time it was a genuine cultural shift. "The captain is not a god" is apparently something you have to teach.
You absolutely have to teach it, and in some cultures you still have to teach it hard. But CRM goes beyond just flattening authority. It includes explicit role division. One pilot flies, one monitors. The monitoring pilot is not just sitting there — they're an independent cross-check. And they swap roles periodically, specifically to prevent what's called monotony blindness. If you stare at the same instruments for hours, your brain stops seeing them.
So let's get into the actual mechanisms. We've established that hypervigilance isn't sustainable and that the industry knows this. What does the cockpit actually do during those six hours of cruise?
The first thing to understand is that the binary model — you're either vigilant or you're relaxed — is wrong. It's not a switch. It's a spectrum, and the optimal point on that spectrum changes depending on the phase of flight. During takeoff and landing, you want high arousal. During cruise, you want what I'd call sustainable monitoring — a state of relaxed alertness where you're not amped up but you're not checked out either.
Relaxed alertness. That's the phrase people use?
It's one of them. The training literature talks about managing arousal states intentionally. The Yerkes-Dodson curve is foundational in pilot training — they actually teach the curve. Pilots learn to recognize where they are on it and to dial up or down as needed.
How do you dial down on purpose at thirty-seven thousand feet?
Part of it is structural. The sterile cockpit rule — FAA regulation fourteen CFR one twenty-one point five forty-two, nineteen eighty-one — prohibits non-essential conversation and activities below ten thousand feet. That's the high-arousal zone. But the flip side, which is less formalized but widely practiced, is that above ten thousand feet during cruise, some conversation, some coffee, some stretching — that's not just permitted, it's understood as part of maintaining sustainable alertness. You're not slacking off. You're managing your cognitive reserves.
Deliberate de-escalation.
Phased de-escalation in awareness. Pilots are trained to consciously lower their arousal state during cruise — not to zone out, but to enter a monitoring mode that can be sustained for hours. The key is that it's paired with periodic re-engagement triggers. The most basic of these is the system scan.
Walk me through a system scan. What does that actually look like?
Structured, scheduled cross-checks of instruments. Every fifteen to twenty minutes, a pilot runs through a specific sequence: altitude, heading, airspeed, fuel state, engine parameters, navigation waypoints, weather radar. It's not a casual glance. It's a deliberate sweep with a known pattern. The pattern matters because it creates a forcing function — your brain can't skip a step if the sequence is baked in.
And the interval matters too, I'd imagine. Fifteen minutes is short enough that nothing catastrophic can develop between scans, but long enough that you're not constantly interrupting the sustainable low-arousal state.
Right. If you scanned every sixty seconds, you'd never settle into cruise mode and you'd burn out. If you scanned every hour, something could go wrong and compound for fifty-nine minutes before you caught it. Fifteen to twenty minutes is the sweet spot that balances safety against cognitive load.
And the co-pilot is doing the same scans, but offset?
Not necessarily synchronized. The monitoring pilot — the one not actively flying — is doing their own independent scans, and they're also watching the flying pilot for signs of fatigue or fixation. CRM creates a second set of eyes that isn't just redundant, it's independent. The monitoring pilot isn't looking at the same things at the same time — they're looking at what the flying pilot might be missing.
That independent cross-check seems like the part that would be hardest to replicate if you're working alone.
And we'll get to that. But first I want to sit with why this system works against both failure pattern. The first failure pattern is forgetting the task — the brain checks out and you stop monitoring entirely. Structured scans and CRM create forcing functions that re-engage attention at regular intervals whether you feel like it or not. The second failure pattern is exhaustion — you try to stay hypervigilant the whole flight and you're fried by the time you need to land. The deliberate relaxation phase preserves cognitive fuel for the high-stakes landing sequence.
The landing is where it all comes together, or falls apart. You've been in low-arousal cruise for hours, and now you have thirty minutes of the most cognitively demanding work in aviation.
And this is where the pre-landing checklist functions as a deliberate arousal ramp-up. It's not just a list of things to verify — it's a psychological transition. You're telling your brain: we're done with cruise, we're entering the critical phase. Verify altitude. Configure flaps. Brief the approach. Review the missed approach procedure. Each step ratchets your attention up one notch, and by the time you're on final approach, you're at peak arousal — exactly where you need to be.
Compare that to how most people prepare for a presentation or a high-stakes meeting. They're either already amped up for hours beforehand and arrive exhausted, or they're distracted and try to snap into focus thirty seconds before.
The pre-presentation ritual for most knowledge workers is ad hoc and often skipped entirely. A pilot would never skip the pre-landing checklist. It's trained, it's standardized, and it's enforced by regulation and by the other person in the cockpit. The structure isn't optional.
So aviation has built this entire apparatus — the Yerkes-Dodson curve, CRM, system scans, sterile cockpit, phased de-escalation, pre-landing ramp-up — all designed around the fact that humans can't sustain peak attention indefinitely. What can the rest of us actually use?
Let's start with the principle that's hardest to internalize: design your environment for the human you are, not the human you wish you were. Most productivity advice assumes you're the human you wish you were. "Just focus for eight hours." "Just don't check your phone." "Just grind."
The hustle advice is basically demanding sustained hypervigilance and calling it virtue.
And it ignores the Yerkes-Dodson curve entirely. The curve says performance drops at high arousal. Sustained high arousal — which is what hustle culture demands — degrades decision-making, narrows attention, and burns you out. Aviation's insight is that the most dangerous state isn't low arousal. It's the wrong arousal at the wrong time. Being relaxed during descent will kill you. Being hypervigilant during cruise will exhaust you before you get to descent. The goal is to match arousal to task demand.
So the first transferable principle is: stop trying to maintain a single peak state. Cycle intentionally.
Cycle intentionally between arousal states. Schedule deliberate low-arousal periods for deep work — not constant high focus, but the kind of sustained attention that's relaxed enough to be sustainable. And then build in forcing functions that re-engage you at intervals.
The Pomodoro Technique is the crude consumer version of a system scan.
It is. Twenty-five minutes of focus, five minutes of break. It's a forcing function — the timer goes off and you're re-engaged whether you drifted or not. But it's missing the cross-check element. There's no second person verifying that you actually did the work during those twenty-five minutes, or that you didn't just stare at the screen and call it focus.
Which is where the solo knowledge worker hits the wall. Aviation's systems work because they're enforced by regulation, by procedure, and by a second human in the cockpit. Translating them to solo work requires self-imposed structure, which is harder.
Harder but not impossible. You can build your own scan patterns. Instead of checking altitude and fuel state, you're checking: have I responded to the thing I was avoiding? Is the document actually progressing? Did I get distracted by something that felt productive but wasn't? A structured self-check every twenty minutes or so.
The pre-landing checklist is the one I keep coming back to. The idea of a deliberate arousal ramp-up before a high-stakes task. Most people just lurch from one context to another and hope their brain catches up.
A pilot's pre-landing checklist takes several minutes and involves both crew members. Verify altitude, configure flaps, brief the approach, review the missed approach. Each step is a cognitive handshake that says: we are entering the critical phase now. A knowledge worker could build a pre-presentation checklist. Review the audience. Confirm the key points. Check the tech setup. Take three deep breaths. It sounds trivial, but the point is that it's standardized and you do it every time, so your brain learns that this sequence means "shift into high arousal now."
And the sterile cockpit rule. No non-essential conversation below ten thousand feet. That maps directly to meeting-free blocks before deadlines. But the rule works because it's enforced by regulation, not willpower. If you're the only one enforcing your own sterile cockpit, you're the captain who can override it whenever you feel like it.
That's the uncomfortable truth underneath all of this. Aviation's Human Factors systems are embedded in an institutional framework with real consequences. The sterile cockpit rule is a federal regulation. CRM is trained and tested and retrained. The second pilot is physically present and has authority. When you're a solo knowledge worker, you're the captain, the co-pilot, and the regulator all in one person. And you're tired, and you know the plane is probably fine, and nobody's watching.
So you skip the scan.
You skip the scan. And most of the time, nothing bad happens. The holes don't line up. Until they do.
The gap between procedure and practice — that's where the real Human Factors challenge lives, not in the design of the checklists.
Procedure is what you write down. Practice is what happens at three in the morning when you're exhausted and you know the system well enough to cut corners and you've cut those corners a hundred times without consequence.
I think this is where Hilbert's been waiting to jump in.
Hilbert: They are.
... Go on.
Hilbert: The systems are good. They're better than what most industries have. But they're designed for ideal conditions and the conditions are never ideal. I worked as a flight dispatcher for a regional cargo airline in the late nineties. Little operation out of Memphis. I planned routes, calculated fuel loads, filed flight plans. But the part of the job that stuck with me was watching pilots come in after red-eye shifts. The company had a rest policy. There was a bunk room. It was next to the baggage conveyor. The noise level made sleep a joke and everyone knew it. The policy existed on paper and nowhere else.
So the rest policy was a compliance checkbox, not an actual rest policy.
Hilbert: It was a room with a door that didn't close all the way and a fluorescent light that flickered. I asked one of the senior captains about it once. He said he'd been flying cargo for twenty-two years and he'd never once felt rested after a bunk-room nap. He just flew tired.
And the scans? The CRM?
Hilbert: Depended on the crew. I knew a captain — good pilot, technically sharp, but he didn't take input. You could see the co-pilots shut down around him. They'd go through the motions of CRM because they had to, but the communication was one-way. The captain talked, the co-pilot confirmed. That's not CRM. That's hierarchy with a different name.
So the culture overrides the procedure.
Hilbert: Every time. The procedure is only as strong as the culture that enforces it. And culture is set by the person in the left seat. If the captain believes in CRM, it works. If the captain thinks it's paperwork, it's paperwork.
You saw people skip scans.
Hilbert: I saw a pilot fall asleep in cruise. Red-eye cargo run, Memphis to somewhere in Texas. The co-pilot noticed — of course he noticed — and he just... let him sleep. Took over monitoring. Didn't log it, didn't report it. Reporting it would have meant both of them pulled from rotation. The co-pilot would have lost hours, lost pay. So he covered for him.
The co-pilot made a rational decision under the incentive structure he was actually operating in.
Hilbert: The plane landed fine. Nothing happened. And that's the problem, isn't it? Nothing happened. The holes didn't line up that night. So the co-pilot learned that covering works, the captain learned that falling asleep has no consequences, and the airline learned nothing because nobody told them.
The system worked in the sense that the plane didn't crash. But the system failed in the sense that the safety mechanisms were bypassed and nobody knows.
Hilbert: I think about that co-pilot sometimes. He did the right thing for the flight — he kept monitoring. He did the wrong thing for the system — he hid the failure. And I don't know what I would have done in his position. He had a mortgage.
This is the part that checklists can't solve. The checklist says "report fatigue events." The incentive structure says "reporting a fatigue event costs you money and your captain will resent you." The checklist loses.
Hilbert: My brother-in-law flies for a major carrier now. Passenger routes. He says it's better than cargo was — more oversight, better rest rules, stronger union. But he also says every pilot he knows has a story about a captain who won't take input, or a scan that got skipped because they were tired and the plane was fine, or a fatigue report that never got filed. The systems are real and they matter. But they're not the whole story.
The whole story is that you're still dealing with humans.
Hilbert: Tired humans in a metal tube at thirty-seven thousand feet who know that ninety-nine times out of a hundred, skipping the scan has no consequence. The system is designed for the one time it does.
And the system mostly works. Aviation's safety record is extraordinary. But Hilbert's point — the gap between procedure and practice — that's where the next layer of work is. It's not enough to design good checklists. You have to design incentive structures that make people actually use them.
Which brings us back to the solo knowledge worker. If aviation — an industry obsessed with redundancy and checklists and formal procedure — still struggles with the gap between what's written down and what actually happens at three in the morning, what hope do the rest of us have?
Less hope, but not zero. The advantage the solo worker has is that they're not fighting institutional inertia. A freelancer can decide tomorrow to build a pre-task checklist and a twenty-minute scan interval and a deliberate wind-down ritual. They don't need FAA approval. The disadvantage is that there's no co-pilot to catch them when they skip it, and no regulator to penalize the skip.
You need to build the co-pilot into the system somehow. An accountability partner, a timer you can't ignore, a log you actually review.
Or you accept that your self-imposed structure will fail sometimes and build that into the design. The Swiss Cheese Model applies to personal productivity too. You don't need a perfect system. You need multiple imperfect layers. A calendar block, a timer, a checklist, a weekly review, a friend who asks how the project is going. None of them is reliable on its own. Stacked together, they catch most of the failures.
There's another layer here that I think we haven't touched. As automation takes over more of cruise — single-pilot cargo flights are already being tested by FedEx and UPS — the Human Factors challenge shifts. It's no longer about managing low arousal in a two-person crew. It's about managing boredom and disengagement in a single pilot watching a system that mostly flies itself.
That's a problem knowledge workers already face with AI tools. When the system does most of the work, your role shifts from operator to monitor. And monitoring is exactly what the human brain is worst at. Sustained, low-demand vigilance without active engagement — that's the cruise problem, but now it's the whole job.
The pilot falls asleep and the co-pilot covers. The knowledge worker lets the AI draft the report and doesn't really read it. Same failure pattern, different altitude.
The solution in aviation will probably be better alertness monitoring — systems that detect when the pilot has checked out and re-engage them. For knowledge workers, the equivalent might be building deliberate review checkpoints into AI-assisted workflows. You don't just accept the output. You run a scan on it — a structured verification pass that forces you to actually engage with what the machine produced.
The pre-landing checklist for AI output.
Verify the claims. Check the sources. Read the thing you're about to send. It sounds obvious, but without a forcing function, it's the first thing to go when you're tired and the output looks plausible.
We've covered a lot of ground. Yerkes-Dodson, Swiss Cheese, Tenerife, CRM, system scans, sterile cockpit, phased de-escalation, the gap between procedure and practice, and what happens when the co-pilot has a mortgage and the captain is asleep. One detail from the research that didn't quite fit but I want to mention: the sterile cockpit rule only applies below ten thousand feet. Above that, conversation is not just allowed, it's understood as part of fatigue management. But the rule doesn't specify what counts as non-essential conversation during cruise. It's left to crew judgment. Which means the boundary between "managing fatigue through conversation" and "distraction" is entirely cultural — it depends on the crew, the airline, the norms of that particular cockpit on that particular night.
Which is Hilbert's point again. The procedure draws the line at ten thousand feet. Everything above that is culture.
Even with all these systems — the scans, the CRM, the regulations, the training, the checklists — the human factor remains the hardest variable. Which leaves us with a question. If aviation can't fully close the gap between procedure and practice, what does that mean for the rest of us trying to apply these lessons without institutional support? Are we building systems we'll actually use, or systems that will look good on paper and fall apart at three in the morning?
I think the honest answer is: some of both. And the difference is whether you design for the human you are or the human you wish you were. The human you are gets tired, cuts corners, has off days. The systems that survive contact with that human are the ones that are easy to do, hard to skip, and have some kind of external accountability baked in.
Thanks to Hilbert Flumingtop for producing, and for the Memphis cargo story that I suspect we'll be thinking about for a while.
This has been My Weird Prompts. If you want to hear more about how weird industries solve universal human problems, subscribe wherever you get your podcasts, and send us your weird prompt at my weird prompts dot com.
We'll be back soon. Try not to fall asleep in cruise.
