On the Specific Discomfort of Discovering You've Been Protecting Something That No Longer Needs Protection

The guy at the auto parts store was trying to help my friend Dave, truly he was, but what he kept doing was pulling out this enormous three-ring binder, the kind with the laminated dividers and the pages in clear plastic sleeves, and running his finger down columns of part numbers while Dave stood there holding a greasy alternator belt that had snapped on his Tahoe somewhere south in the middle of Nevada on what was supposed to be a quick road trip to Vegas.¹ I was leaning against a rack of air fresheners, not really involved, which is maybe why I noticed what I noticed: this man spent eleven minutes doing something Dave had already done on his phone in the parking lot in about forty-five seconds. He had the part. He knew he had the part. The binder was a ritual, not a search. And the thing is, he wasn't doing anything wrong. The binder had been, for twenty years, the correct tool. The binder had kept this store running, had prevented mis-orders, had served as a physical record of every compatible gasket and filter and tensioner pulley in the Bosch catalog. But the binder was now a skeuomorph, a residual artifact of a constraint that no longer existed, and this man was spending the most valuable minutes of the interaction performing obeisance to it.

I have been thinking about that binder a lot lately. Not because I care particularly about auto parts retail (though I find it weirdly soothing), but because I have lately become convinced that I, personally, am full of binders. That my entire professional life is structured around habits and rituals and processes that exist to protect against a scarcity that is, right now, in the process of evaporating so quickly it's almost impossible to feel the temperature change while you're standing in it.²

Here is what I mean. For most of the time any of us have been working, the single most expensive thing in any knowledge-work organization has been execution. Building things required people with rare skills, and those people took years to train, and their hours were precious, and so we evolved an elaborate set of rituals to make sure we didn't waste those hours on the wrong problems.³ Planning documents. Alignment meetings. Approval chains. Requirements gathering. Stakeholder reviews. Consensus-building exercises so thorough they sometimes felt designed less to achieve agreement than to distribute blame should the thing we built turn out to be wrong. Which, of course, it usually was, at least partly, because plans have this stubborn tendency not to survive contact with the actual world they're meant to describe.⁴

All of which made sense. It made perfect sense. If the meeting to discuss whether to build a feature costs six person-hours, and the feature itself costs six person-months, you definitely hold the meeting. You hold several meetings. You wordsmith the product requirements document until it gleams with the cold perfection of something nobody will enjoy reading. Because the alternative, building the wrong thing, is catastrophic. You've just burned a quarter of engineering time on a direction the customer didn't want. So we ritualized. We systematized the pre-work. And over time (this is the part that matters), the rituals stopped feeling optional. They calcified into the way work is done, full stop, no further justification needed, sort of how your grandmother's recipe calls for sifting the flour three times because that's what you do with flour, never mind that modern milling has made triple-sifting roughly as useful as churning your own butter.⁵

William James, writing in 1890, understood something about this. He argued that habits, once formed, become automated pathways in the brain, little neurological grooves that allow us to act without conscious deliberation. This is, on balance, a spectacular feature of being human. The more of our daily operations we can hand over to automatic processes, the more cognitive bandwidth we free up for actual thinking, for the kind of work requiring genuine attention and judgment. But James also recognized the dark complement to this efficiency: the groove, once cut, is hard to reroute. We become, in his famous phrase, "mere walking bundles of habits," and the bundles persist long after the conditions producing them have changed.⁶ The habit no longer serves the function it was designed for, but it keeps firing anyway, burning energy, consuming time, occupying the slot where a better response could live.

This is, I think, what is happening right now to a very large number of people in a very large number of organizations, and also (I keep catching myself at this) to me specifically. The constraint around execution capacity is dissolving. Not everywhere, not for every kind of work, and not completely.⁷ But measurably, observably, in ways that are producing genuine disorientation for anyone who built their career around the implicit understanding that building things is hard and slow and expensive and must therefore be protected at all costs from misdirection. And the disorientation isn't because the old constraint disappeared. It's because the constraint moved, and we didn't move with it. We're still standing at the counter, running our fingers down the binder.

There's a principle in manufacturing, originally articulated by Eliyahu Goldratt in his Theory of Constraints, that goes something roughly like this: when you eliminate a bottleneck in a system, the bottleneck doesn't vanish. It migrates. It shows up somewhere else downstream, often in a place you weren't looking, because you had no reason to look there when the old bottleneck was dominating everything. The resistance isn't destroyed; it's relocated. And if you don't go looking for the new bottleneck, you'll spend all your newly liberated capacity optimizing a part of the system that no longer needs optimizing while the actual constraint, the real chokepoint, quietly compounds.⁸

So where did it go? Where is the bottleneck now, if it's no longer in execution?

I think it went to roughly four places, and the reason I find this worth writing about at length (aside from the fact that I find most things worth writing about at length) is that these four places correspond to capacities we have been systematically underinvesting in, both as individuals and as organizations, because they were never the binding constraint before. They were nice-to-haves. Soft skills. The kind of thing your company sends you to a half-day workshop about and then never mentions again.

The first is clarity. Knowing what's worth building. Which sounds obvious until you realize that for most of corporate history, we used planning processes as a proxy for clarity. The PRD, the roadmap, the quarterly planning cycle; these weren't just coordination tools, they were forcing functions for thought. They made you sit down and articulate what you wanted before you spent the engineering budget. But they were always a substitute for the real thing, and the real thing (actually understanding what the customer needs, what problem you're solving, whether the problem is even the right problem) was always the hard part. Now that you can build a working prototype in an afternoon, the fact that you don't know what to build becomes visible almost immediately. The lack of clarity, previously hidden behind six weeks of planning ceremony, is exposed in the first hour.⁹ And it turns out most of us, myself very much included, are less clear on what we actually think than we assumed we were. The planning process was doing some of our thinking for us, and without it, there's a kind of cognitive vertigo, a moment where you realize you've been leaning against a wall that someone just removed.

The second is ambition. Or rather, the willingness to swing for something large. When execution is expensive, conservatism is rational. You get three or four shots per year, maybe, to build something meaningful. So you hedge. You build the incremental improvement, the safe bet, the feature the loudest customer asked for. And you call this being data-driven. Or customer-focused. Or pragmatic. But when you can take fifty swings a year? The risk profile inverts completely. Timidity becomes the danger. Building small when you could build transformatively is, in this new math, a form of waste as real as building the wrong thing ever was.¹⁰ And I notice, in myself and in people around me, a deep reluctance to think at the scale the tools now permit. We keep building what amount to, and there's really no kinder way to say this, horseless carriages. Products and processes conceived entirely within the mental model of the old constraint, dressed up in AI trim but fundamentally shaped by the assumption that execution is precious and must not be risked on anything too ambitious.¹¹

The third is distribution. When everyone can build, the product itself stops being the competitive advantage it once was. Getting the thing into people's hands; establishing the relationships and channels and trust networks necessary to reach actual users at scale; that's the hard part now. I know engineers (and I count myself adjacent to this tribe) who find this almost offensive, the idea that the beautiful thing you built matters less than your ability to get it in front of someone. But it's been quietly true for a long time, and the AI acceleration of execution has just made it undeniable. There is a saying in startup circles about how most founders overinvest in product and underinvest in go-to-market. That ratio is now more lopsided than it has ever been.¹²

The fourth is relationships. And I realize this sounds soft, sounds like the kind of thing a career coach would say while gesturing at a whiteboard with "TRUST" written on it in green marker. But the argument is structural, not sentimental. When technical capabilities are compounding on a near-weekly basis, when the platform you mastered last quarter might not be the right platform next quarter, the durable thing, the thing with a half-life longer than an LLM release cycle, is the human web of mutual reliability and trust you've woven around yourself.¹³ You can't prompt your way into a reputation for following through. You can't automate the judgment call about which person to call when something goes sideways. These are relationship assets, and they compound in the same way, just more slowly and less visibly, as the technical assets everyone is focused on.

Here's what I keep catching myself doing, though, and this is the part where the essay has to get uncomfortably personal or it's not worth writing. I know all of this. I've just articulated it with enough coherence to suggest I at least partially understand the landscape. And yet last Tuesday I spent ninety minutes preparing a slide deck for an internal meeting about a process I could have prototyped in less time than it took me to find a good stock photo for slide three.¹⁴ I spent ninety minutes performing the ritual. The binder. The thing I know how to do, the thing the groove in my brain says is responsible and thorough and professional. And I knew, while I was doing it, that I was doing the equivalent of sifting the flour for the third time. I just couldn't stop. Or didn't stop. There's a difference, but it's smaller than I'd prefer.

This is, I think, the real difficulty. Not the intellectual recognition that constraints have shifted. That part's almost easy; you read one article, you see one team ship a product in ten days with four people (Claude Cowork), and the logic is apparent. The hard part is behavioral. The habits are deep. They've been rewarded for years. The permission loop (checking before doing, confirming before building, waiting for the Slack reply before moving forward) feels responsible. It feels like diligence. Skipping it feels like arrogance, or recklessness, or that particular kind of professional negligence your manager will eventually notice in a performance review.

Goodhart's Law tells us that when a measure becomes a target, it ceases to be a good measure. Something similar happens with work rituals: when a process designed to reduce risk becomes the thing you do to feel safe, it stops reducing risk and starts producing drag. The feeling of safety is skeuomorphic; it resembles the real thing without providing any of its structural benefits.¹⁵ You're still exposed to the same risks (wrong direction, misunderstood customer, wasted time), but you've also added the cost of the ritual itself. You've paid for insurance that doesn't cover anything.

And the worst part (or maybe the most interesting part, depending on your tolerance for watching yourself fail) is that the rituals are self-reinforcing. You prepare the deck because the meeting expects a deck. The meeting exists because the process says align before you build. The process exists because someone, probably years ago, watched a team build the wrong thing and decided the solution was more pre-work. Each layer justifies the others. It's turtles all the way down, except the turtles are calendar invites.

I should say: I'm not arguing for chaos. I'm not arguing that thinking is a waste of time or that planning is obsolete.¹⁶ Good thinking is, if anything, more valuable now than ever, precisely because the gap between having an idea and having a thing in the world has collapsed to almost nothing. What gets built will increasingly reflect the quality of thought behind it, without the planning process acting as a filter or refinement stage. If your initial idea is vague and unexamined, the prototype you produce in an afternoon will be vague and unexamined, and you'll have spent an afternoon learning that you hadn't thought hard enough, which is useful but expensive in a way that's different from the old expense. The old way wasted execution. The new way wastes iterations. Both are real costs.

But the balance point has shifted, and I think the shift requires something from us that is genuinely uncomfortable, which is a willingness to be seen mid-process. To show the rough version. To circulate the half-formed thought. To let people see you before you've figured it out completely. There's an ego cost to this that nobody talks about enough. We polish not just because polish was once protective but because polish is a form of self-presentation. The finished deck says: I am competent. I am thorough. I have anticipated your questions. The rough prototype says: I had an idea and I'm not sure it's right and I'd rather find out now than in six weeks. The second statement is more honest, more efficient, and far more vulnerable. And vulnerability, as it turns out, is not something we were trained for in most professional environments. We were trained for competence. We were trained for the binder.¹⁷

What I find myself wanting to do (and this is the part I don't have an answer for, just an observation that feels true enough to be worth recording) is to build a practice, slowly and badly and in public, of noticing when I'm performing the ritual versus when I'm doing the work. They look almost identical from the outside. Sometimes they look identical from the inside. The difference is in what they're protecting. The ritual protects my sense of being a responsible professional. The work protects the thing I'm actually trying to accomplish, which is sometimes a product and sometimes a relationship and sometimes just a clear answer to a question I should have asked three weeks ago.

I think about that guy at the auto parts store. He knew the part number. He could feel it in his fingers, I'm almost sure of it, that particular knowledge you develop after two decades of matching serpentine belts to engine codes. But the binder was there, and the binder was the process, and the process was what it meant to do the job correctly. And I understand that. I understand it so well it makes my teeth hurt, even though I was just the guy leaning on the air freshener rack watching it happen. Because I am also standing at that counter, every day, running my finger down the column, confirming what I already know, in a language that used to be the only language available and is now just one of several, and not the fastest one.

The belt, meanwhile, doesn't care about the binder. The belt just needs to be replaced.¹⁸

¹ The Tahoe, for what it's worth, has the kind of mileage that makes mechanics wince, and Dave refuses to replace it on principle. He says it has character. I say it has symptoms. We were about ninety minutes into the drive when the belt went, which meant we were too far from Reno to turn back and too far from Vegas to care about anything except finding the part. ↩︎

² I recognize this sentence is doing a lot of metaphorical work. Temperature change for scarcity shift. Standing for passive acceptance. I'm not going to apologize for it but I wanted you to know I noticed. ↩︎

³ "Rare skills" is doing some heavy lifting here, and I don't want to imply that skill has become irrelevant. It hasn't. What's changed is the specific set of skills that constitutes a bottleneck. The skills required to be a good thinker, a good communicator, a good judge of what to build: these are as scarce as they ever were, possibly more so. ↩︎

⁴ This observation has been attributed to so many military figures (Eisenhower, Moltke, Mike Tyson if you count "everyone has a plan until they get punched in the mouth") that its provenance has become its own kind of object lesson about how ideas survive by being useful, not by being correctly attributed. ↩︎

⁵ Though, and I feel compelled to add this, there is apparently a school of baking that insists triple-sifting does produce a meaningfully lighter crumb structure. I am not qualified to adjudicate this. ↩︎

⁶ James wrote this in The Principles of Psychology, which is itself a kind of meta-example of the principle it describes: the book's influence grooved so deeply into the field of psychology that its frameworks are still shaping how we talk about mental processes 135 years later, long after the neurological models it proposed have been superseded. ↩︎

⁷ A caveat worth more space than a footnote, probably. There are fields, medicine and law and structural engineering among them, where execution remains genuinely expensive and the consequences of getting it wrong remain genuinely severe. The question in those fields isn't whether the constraints have shifted (some of them have, some haven't) but whether the rituals surrounding the work are calibrated to the actual risk or to a historical risk level that no longer applies. My suspicion is it's both, in varying proportions, and the proportions are worth examining case by case. ↩︎

⁸ Goldratt, incidentally, communicated this idea primarily through a novel about a struggling manufacturing plant manager, which is itself a kind of argument for the primacy of storytelling over abstraction in getting people to change their behavior. The theory is rigorous. The delivery vehicle is fiction. Make of that what you will. ↩︎

⁹ This is, I think, why so many people report feeling simultaneously empowered and terrified by AI tools. The tools remove the lag between intention and artifact. Which is fantastic if your intention is clear and well-considered, and acutely uncomfortable if it isn't. ↩︎

¹⁰ I want to be careful here. There's a difference between ambitious thinking and reckless thinking, and the line between them is not always obvious, especially from the inside. The point isn't to abandon judgment. It's to recognize that the frame within which we exercise judgment has expanded, and that exercising the same level of caution within a larger frame produces a proportionally smaller outcome. ↩︎

¹¹ The term "horseless carriage" is itself a perfect example of what it describes: a new thing named entirely in terms of the old thing it's replacing, because the people encountering it lacked any other frame of reference. We called them horseless carriages until we had enough experience with cars to call them cars. The question is how long that renaming takes, and whether you can shorten it by noticing you're doing it. ↩︎

¹² The Cognition/Infosys partnership is instructive here. Cognition, makers of the AI coding agent Devin, chose to partner with Infosys and deploy through their global client base rather than build their own distribution from scratch. The technology was the easy part. The enterprise relationships were the moat. ↩︎

¹³ This is one of those observations that is simultaneously obvious and strangely difficult to act on. Everyone agrees that relationships matter. Almost nobody allocates their time accordingly, because relationship-building doesn't produce the same immediate, visible output as building a feature or writing a document. The return is real but lagging and diffuse, which makes it almost invisible to any system optimized for short-term throughput. ↩︎

¹⁴ The stock photo was of a handshake. It is always of a handshake. ↩︎

¹⁵ I realize I'm using "skeuomorphic" twice in one essay, which is either thematic consistency or intellectual laziness. Possibly both. ↩︎

¹⁶ Though I will say that I've watched planning cycles in large companies take longer than some teams now take to ship an entire product, and the irony of that is so thick you could laminate it and file it in a three-ring binder. ↩︎

¹⁷ There's something worth noting here about the generational dimension of this shift. People who are early in their careers, who haven't yet had twenty years for the old grooves to set, might find the transition easier. Not because they're smarter or more adaptable in some innate sense, but because they have fewer habits to unlearn. James observed that the period before thirty is when habits form most readily, and by extension, the period after thirty is when they're hardest to dislodge. I'm well past thirty. I can feel the grooves. ↩︎

¹⁸ Dave and I made it to Vegas about four hours late. On the drive back a few days later, we stopped at the same store because Dave needed washer fluid. Different clerk. He scanned the barcode on the empty jug with his phone, pointed Dave toward the shelf, and the whole thing took about ninety seconds. The binder was still on the counter. Nobody was using it. ↩︎