The smell of burnt dust from the HVAC vents always hits the back of my throat right before the quarterly performance review. It is a dry, metallic scent that suggests something is working too hard to keep the atmosphere breathable.
I was sitting in a room that smelled like that , listening to a lead architect explain a new “Standardization Suite” that would automatically block any code or process that didn’t meet the company’s defined best practices.
While he was mid-sentence-talking about “algorithmic guardrails” and “eliminating human variance”-I yawned. It wasn’t a small, polite yawn. It was a deep, jaw-cracking betrayal of my own body. He stopped. The silence in the room became heavy, the kind of heavy you only feel when you’ve accidentally insulted a man’s religion.
The Religion of Frictionless Excellence
But standardization is a religion. It is the belief that if we can just make everyone act like the average of the top ten percent, we will achieve a state of permanent, frictionless excellence. It is a lie, of course.
Consider the desktop stapler. It is a handheld machine for the delivery of specific, violent conformity. It requires a singular gauge of wire, a specific arc of tension, and a flat surface. If you try to bind sixty pages with a stapler rated for fifty, the machine doesn’t merely decline the task. It turns the metal staple into a crumpled, useless knot that prevents the next fifty “correct” actions from occurring.
The stapler is a system designed to ignore the thickness of the story it is holding. It only cares about its own capacity.
When we build tooling to enforce best practices, we are building digital staplers. We are saying that the “thickness” of the problem-the nuance, the edge case, the brilliant but non-standard workaround-doesn’t matter. Only the gauge of the wire matters.
Sarah’s Bottleneck: When Protection Becomes Poison
Last month, a developer I’ve worked with for years, let’s call her Sarah, hit the wall. She was dealing with a legacy database migration that was throwing errors no one had seen since . She found a way to bypass the standard handshake protocol, a clever bit of “wrong” code that would have resolved the bottleneck in four minutes.
But the new enforcement tool wouldn’t let her commit the change. It flagged her solution as a “violation of architectural patterns.” The tool was right: it was a violation. But the tool was also stupid: it didn’t know that the “correct” pattern was exactly what was causing the database to choke. The tooling that was meant to protect the system’s quality ended up mandating its failure.
The Hospice Paradox
In my work as a hospice volunteer coordinator, I see this same tension between the “approved way” and the “better way” play out in a much more visceral arena. There is a “best practice” for delivering difficult news. There is a “best practice” for managing a family’s expectations. These are often codified into checklists that volunteers are supposed to follow to ensure “uniformity of care.”
But you cannot script a goodbye. I once watched a new volunteer-a man who lived and breathed by the manual-try to comfort a woman whose husband was in his final hours. The manual said to “offer a glass of water and maintain a respectful distance.” He did exactly that. He was a perfect avatar of best practice.
“The woman didn’t want a glass of water. She wanted someone to sit on the floor with her and talk about the time her husband accidentally dyed their cat blue.”
– Narrative Observation
The “best practice” was a barrier to the “better practice.” The better practice required a deviation from the norm, a messy, non-standard human moment that no manual would ever dare suggest.
We see this same dynamic in the digital space. Companies like
understand that while automation handles the bulk of the work-the deposits, the withdrawals, the 24/7 access-there must be a layer of human judgment that isn’t suffocated by the machine. In a highly regulated, high-speed environment, the “best practice” is transparency and speed, but the “better practice” is having a professional team available to handle the weird, the glitchy, and the human. Consistency is the floor, not the ceiling.
The Hidden Cost of Automated Enforcement
81% of safety stops are triggered by high-performing operators trying to optimize a flow sensors can’t yet recognize.
In plain human terms: eight out of ten times the “guardrail” screams at you, it’s not because you’re failing; it’s because you’re trying to be better than the machine is allowed to be.
This is the hidden cost of automated enforcement. We aren’t just catching mistakes; we are outlawing expertise. When you tell a team of seniors that they can only use the patterns that a junior can understand, you aren’t raising the floor. You are lowering the ceiling until it’s a crawl space.
Consistency is the lowest form of quality. It is a baseline. It is the assurance that things won’t be terrible. But excellence? Excellence is almost always an outlier. It is the deviation that works. If you build a system that forbids deviation, you have successfully built a system that forbids excellence.
The Ghost in the Guardrail
The architect who I yawned at eventually finished his presentation. He looked at me and asked if I had any “constructive feedback” regarding the new enforcement suite.
I thought about Sarah and her database. I thought about the volunteer and the glass of water. I thought about how many hours we spend at our desks trying to trick our tools into letting us do our jobs well.
“The tool is perfect,” I said. “It will ensure that we never make a creative mistake again. It will also ensure we never find a creative solution. We are building a cathedral out of nothing but 90-degree angles, and then we’re going to wonder why it feels so cold inside.”
He didn’t like that. He wanted to talk about “scalability.” But scalability is often just a polite word for “making things simple enough that we don’t need to trust the people we hire.”
When you trust the tool more than the person, you’ve already lost. The tool can only see what has been programmed into it-the past. The person is the only one who can see the present. The “best practice” is a map of where people have already walked. The “better practice” is the path you have to cut yourself because the map is out of date.
The Wisdom of the Wise Deviation
We need a new way to think about our guardrails. Instead of “enforcement,” we should think about “advisory.” A tool that says, “Hey, this is unusual, are you sure?” is a powerful ally. A tool that says, “I am locking your keyboard because you aren’t doing it the way I was told to do it,” is an adversary.
The best work happens in the friction between the standard and the exception. It happens when an experienced professional looks at the “best practice” and says, “Not today. Today, the context demands something different.”
If we continue to automate the judgment out of our systems, we will end up with a world that is perfectly consistent and entirely mediocre. We will have staplers that never jam because we’ve stopped trying to bind anything thicker than a pamphlet. We will have code that never breaks because we’ve stopped trying to solve the hard problems.
I’m still waiting for the day when a tool is smart enough to recognize a “wise deviation.” Until then, I’ll keep yawning in meetings where people mistake consistency for quality. I’ll keep cheering for the Sarahs of the world who find the cracks in the guardrails. Because that’s where the light gets in-through the violations of the pattern.
The staple that insists on a perfect bind is the same tool that leaves the sixty-first page drifting on the floor.
We should be building platforms that empower the human at the center of the experience, rather than tools that treat human judgment as a bug to be patched out of the system. Whether it’s in the way we care for the dying or the way we build an entertainment platform, the goal shouldn’t be to eliminate the outlier. The goal should be to give the outlier a place to work its magic.