The Death March Is a Confession, Not a Strategy

Let’s be clear. I like to go fast. Not in a metaphorical, spiritually-impoverished-modern-man sort of way. Fast. I race motorcycles. For pleasure. Which is, I’m aware, the medically-recognized definition of a death wish with a hobbyist’s budget. I ride a liter bike—a Ducati, under ten thousand miles—that has, in that admittedly brief tenure, consumed four sets of rears, two fronts, and a set of brake pads. That’s not maintenance. That’s an eating disorder with a warranty. And I want to be exact about the appeal, because the exactness is the appeal: no one’s forcing me. There’s no man on a pit wall, screaming. There’s no deadline. There’s just the elegance—and I use the word in its full, load-bearing, aesthetic sense—of getting from A to B at the maximum velocity I can, and I’d underline this word if the format allowed, control. I love the precision. I love the flow state, that condition in which the self briefly, mercifully, shuts up. And most of all I love the one thing racing has that meetings never will: the complete absence of bullshit.

But here’s the thing. When I turn up to a circuit I’ve never ridden, I don’t invent a lap time and then attempt to achieve it via the medium of manslaughter. I don’t announce, “I’ve decided the optimal lap is 1:32 flat, and I’ll be firing my own braking hand if we come up short.” Because if I did, any rider with a functioning brainstem would deposit me gently into a gravel trap and summon an ambulance, and they would be correct to do so. And yet. That is, with almost no exaggeration, the software death march. We’ve simply let the people who have never turned a wrench, or written a line of code, or shipped so much as a working toaster, rebrand it as “visionary leadership.” Bold vision. Someone else’s funeral. Where do I sign?

When I build something for myself, I don’t import every process I’ve ever suffered through at work. I use the minimum viable amount of ceremony, because I am an adult with a finite lifespan. When I hang a picture, I don’t relocate the entire garage into the living room and arrange it by descending tool size on the off chance the drywall demands a torque wrench. That would be the behavior of a man who needs supervision. But the death march assumes precisely this—that everyone in the building is an idiot except the visionary who picked the date. And let’s be honest about where dates come from. No engineer has ever volunteered one. Not once. Not in recorded history. You ask us for an estimate—a probabilistic range, with error bars, offered in good faith—and you take the cheerful end of it, delete the error bars, weld it to a quarterly earnings narrative, and call it a “commitment.” That’s not a plan. That’s a hostage video with a Gantt chart. Question: if the number was always going to be non-negotiable, why did you ask?

That’s not a plan. That’s a hostage video with a Gantt chart.

So how did we get here? Trust, mostly. Specifically the absence of it. We got here because a particular managerial caste—people whose core competency is converting ambiguity into slides—concluded that squeezing humans is a more scalable skill than making anything a human would want. We got here because someone decided that if I’m not visibly typing every second of the day, I must be stealing from the company. As though thought were a batch job you could schedule between stand-ups. As though the good ideas arrive on a cron. And there is a lovely phrase this caste enjoys—“human capital”—which reduces the entire catastrophe of a human life to a line item you can depreciate. Hannah Arendt spotted the move some time ago: turning people into interchangeable units of labor is the necessary first step to treating them as disposable. She wasn’t writing about Q3. She might as well have been. Now—does putting a gun to someone’s head make them work faster? It does. For a while. Right up until the part they don’t put in the productivity deck.

American labor has never been more productive. That’s not a slogan, it’s a chart, and the chart goes up and to the right in exactly the manner executives claim to adore. So here is a genuinely confusing question: where did the money go? Productivity went vertical. Wages went for a lie-down. Rent became a personality trait. Healthcare became a GoFundMe. And the official explanation is that we simply haven’t been trying hard enough—which is a bold thing to say to the most productive workforce in human history. The surplus got extracted, the remainder got financialized, and then—this is my favorite part—we were asked to be grateful for the pizza. Thin crust. Lukewarm. In lieu of the money. And we said thank you. Why did we say thank you?

Most of my actual breakthroughs—the real architectural insights, the genuine product ideas—did not happen at the desk. They happened on the walk to the coffee shop, next to a colleague smarter than me, while we talked about nothing in particular and let our brains idle. Not interrogating each other. Not “syncing.” Just walking. The mind needs slack the way an engine needs an oil film: remove it in the name of efficiency and you get efficient, spectacular, metal-on-metal failure. This is very difficult to reproduce under duress. It is genuinely impossible to reproduce under the chronic, ambient threat of the next reorganization—the RIF, the “workforce optimization,” whatever the comms team has focus-grouped this quarter and announced, as tradition demands, alongside record profits. So the plan is: manufacture constant fear, then sit back and await the calm, original, high-trust thinking. How’s that going?

Why does it keep getting worse? Because until every waking second of my life has been converted into shareholder value, the appetite will not be satisfied—and, spoiler, it will not be satisfied then either. Billionaires are, clinically, a hoarding disorder that society elected to admire rather than treat. We dress it up—grit, hustle, the sanctity of shareholder value—but strip the branding off and stare at the raw quantities. If a man owned a billion shoes, we’d stage an intervention. A billion cars. A billion boats. A billion of anything is a diagnosis, not an achievement. And at that scale something goes in the wiring: you become genuinely unable to perceive the people you employ, because the sheer distance has sanded every human detail off them. They’re not evil, necessarily. They’re just very, very far away. Which is the uncomfortable question, isn’t it—if you truly can’t see the person anymore, is it still contempt, or is it just altitude?

I started in tech in 2010. The tooling since then has become genuinely miraculous. Because I stand on the shoulders of giants, I can now build in a month what used to take twenty or thirty people a full quarter to plan and ship—frameworks, cloud infrastructure, the entire open-source ecosystem quietly doing the work of an army. Objectively, we go faster every year. And the response to this abundance is: not fast enough. It has to be a death march. It has to hurt. And the death march does real, measurable damage—it wrecks teams, corrodes marriages, hospitalizes the occasional nervous system (this isn’t hyperbole, I have the receipts)—all of which gets filed under “the cost of doing business” by people who are, notably, not the ones paying it. David Hume said reason is the slave of the passions, and he wasn’t being unkind, he was being accurate. These men aren’t reasoning toward the deadline. They’re rationalizing an appetite they’ve declined to examine. Which raises a small arithmetic problem: if the tools got ten times faster and the pressure somehow got worse, what exactly was the pressure ever about?

Reason is, and ought only to be, the slave of the passions.

Despite being a demonstrably losing strategy, we do more of it every year. Why? Because an arbitrary deadline is a marvelously efficient way of telling your staff, without technically saying it, “you are replaceable, and I can find a faster grunt.” And with AI in the room, leadership is now betting the entire farm on that precise assumption. So I’ll say the quiet part at conversational volume: I think in another ten years I’ll operate at maybe ten times my current output—some of that AI, most of it the rest of the tooling stack finally growing up. AI is one ingredient. It is not the meal. But a certain kind of manager has watched a language model produce something shiny and concluded it produced something finished. Those are not the same word. I would know—I’m a software architect at a five-thousand-person enterprise SaaS company in identity and access management. We build the pipes through which every employee reaches every system. A hallucinated access policy, a subtly wrong token validation, an AI-generated config that “looks right” and quietly opens a privilege-escalation path—these are not, as the roadmap insists, “iteration opportunities.” They are the incidents that make the news with your company’s name in the headline. So forgive me if I decline to YOLO unreviewed code into a production IAM system to hit a date someone invented in a spreadsheet. Serious question: which slide covers the breach?

Why doesn’t this work for leadership? Let’s take it in order, because it’s a rich text. You get high turnover—though you may not mind, since to you churn is merely another metric with a pleasing downward option. You get an exhausted, less effective workforce, because you cannot death-march a thinking team into sustained excellence any more than you can hold a motor at redline and expect to keep the motor. You paint a target on your own back, because eventually a hungrier suit notices you’ve built no real relationships and lifts your job with the ease of picking up dropped keys. You leave with a severance package and a reputation for hitting dates nobody believed in. And you erode the customer relationship one broken promise at a time, until “ambitious roadmap” and “serial fabulist” become, in their minds, the same phrase. Then you land at the next company and do it all again. Genuine question: at which point in this plan did you win? Or, do you think if you embrace a pathology of lying, you’ll also inherit Musk’s status—even without an emerald mine as a head start? Just another temporarily displaced trillionaire, see.

Why doesn’t it work for customers? Because you ship them something half-built and call it a launch. Corners get cut to make the date, so the features that made the thing worth using quietly fail to make it into the box. Customers get angry. And here is the actual tragedy of it: even when the underlying idea was genuinely good, a rushed implementation kills it in the cradle. That’s a false negative—you’ve killed a winning idea and concluded the idea was the problem. You’ve left money on the table, possibly a great deal of it, to hit a date that impressed exactly one person in exactly one meeting. Then the technical debt compounds until every step forward requires two steps back, engineering resents sales, sales resents engineering, and everyone sprints flat out to remain precisely where they already were. All of this to honor a number you made up. Was it, on reflection, a good number?

Why doesn’t it work for the people actually building the thing? The good ones leave, and the ones who stay are, with total candor, staying for the paycheck—the intrinsic motivation has been surgically removed and replaced with transactional compliance. Standards fall, because you shipped a hackathon prototype to close a sale, and now it needs six months of rework just to remain upright in production. The architecture becomes a patch on a patch on a patch, load-bearing in ways nobody dares to document. Uptime degrades for reasons that grow steadily more specific and more haunting, and because you’re not an engineer and regard the sausage-making as beneath you, you file every outage postmortem under “excuses” and fire somebody, which reliably makes it worse. The senior people go. The juniors inherit systems they were never meant to hold alone, at triple duty, and it breaks—and it is not their fault. It is the wholly predictable result of treating expertise as interchangeable and talent as infinite. Both of which are, and I cannot stress this enough, finite. Who, exactly, told you they weren’t? Was it your friend Bezos? How was his wedding?

So what’s the fix? It’s almost insultingly simple. Let us do the fucking job that we already want to do. Trust us when we tell you we’re going as fast as we responsibly can. Respect labor and stop treating humans as replaceable cogs in a machine. Treat layoffs like you’re deciding which of your children you’re going to starve because when you lay people off you put theirs at risk. Get in the room and build alongside us, or trust the people you’ve already put in the room to carry your urgency without setting the building on fire. We want to win too. We like money. We would simply prefer that your success stop arriving as an itemized bill at the expense of our health in a country where healthcare isn’t a given. Because we’ve noticed the cycle: you gamble our wellbeing, the stock ticks up five percent, you collect a bonus with more commas than a Russian novel, and then—as a thank-you for the quarter—you lay off a tenth of us. We see it. We see the contempt idling quietly beneath the “we’re a family” email. Quick question, since we’re apparently family: which of your actual relatives do you treat like this?

Émile Durkheim warned that when economic goals get unhooked from any social limit, they turn infinite and insatiable—he called it the “malady of infinite aspiration.” That’s the diagnosis. Not ambition. Ambition has a destination. This is a treadmill with the stop button removed, and you’ve decided to make it everyone’s problem but your own. So, honestly: when was the last time enough was enough? Take your time. I’ll wait.

I ride fast because I understand the machine, I respect the track, and I have never once confused recklessness with speed. That last distinction is the entire job, and it’s the one the death-march class has somehow never learned. But the track doesn’t lie, and neither does the product. The tide goes out. The crashes arrive on schedule. And when the noise finally clears, the only people still upright are the ones who built something real, together, at a pace that didn’t require destroying everyone to sustain it. That isn’t idealism. That’s just the lap time. So here’s the last question, and it’s a sincere one: knowing all of this—knowing precisely how it ends, every single time—why are you still holding the stopwatch?