Simplification: Not As Simple As It Sounds
The systems we find most confusing or frustrating are the same ones responsible for our standard of living.
The global supply chains that feel opaque and unaccountable. The financial structures that even the “experts” struggle to explain. The regulatory frameworks that seem to grow by accretion with no evident design. The distributed labor markets that feel arbitrary in ways that seem detached from simple logic. These are the exact mechanisms behind decades of incremental improvements that only make sense if we had the ability to see the entire ecosystem of human interactions and inherent incentive structures. It’s only when we zoom out that we observe the macro: changes in human outcomes so dramatic that earlier generations would have considered them miraculous. Child mortality in wealthy countries fell by more than 90 percent over a hundred years. Life expectancy in those same countries went from roughly 50 to nearly 80. A billion people moved out of extreme poverty within a single generation. Global trade made caloric sufficiency a baseline for most of the world rather than a privilege for the few.
None of this happened despite the complexity. It happened through it.
What the numbers represent
As recently as a century ago, production and consumption were largely local. If the regional harvest failed, people starved. If the local economy contracted, there was no distant system to absorb the shock. The material constraints of your geography were the immediate constraints of your life.
The systems that loosened those constraints are the same systems that made modern life difficult to read. Global supply chains required coordination between parties who never meet, operating across jurisdictions with different legal systems, responding to price signals produced by markets no single participant fully understands. Cross-border financial flows required abstraction layers between lenders and borrowers that stripped out the personal relationship while enabling transactions at scales no personal relationship could reach. Labor markets spanning entire continents required information infrastructure and legal frameworks that no single actor designed or maintains.
The layering, the abstraction, the long chains of causation that make modern systems so hard to trace: these are the features that allow the outcomes to occur at all. The coordination that produces the outcomes requires the same illegibility that produces the confusion. You are not dealing with a good machine that also has an annoying design flaw. You are dealing with one thing, encountered from two different positions in the chain.
An example we wear every day
A cotton shirt.
The fiber was grown in one country, ginned in another, spun into thread, woven into fabric, cut and assembled in a facility elsewhere, shipped through a logistics network spanning multiple ports, tracked through customs in several jurisdictions, warehoused, loaded onto a truck, and carried to your door. The price you paid reflects cotton market conditions, currency exchange rates, fuel costs, labor regulations in multiple countries, platform fees, insurance, financing, and last-mile delivery economics.
No single person designed this chain. Nobody runs it from the center. The buyer, the manufacturer, the carrier, the customs authority, the logistics software firm: each operates within constraints shaped by the others, responding to their own information and incentives. None of them single-handedly designed the entire ecosystem. The system produces an outcome that none of its participants fully controls.
And yet it reliably delivers the shirt. For less than the cost of a meal.
A comparable garment, produced locally from local materials by skilled labor with no global supply chain, would cost multiples more and be available to far fewer people. The complexity is not a tax on the product. It is the mechanism that makes the product possible at the price it arrives.
We almost never register this as remarkable.
How complexity accumulates
The fundamental driver behind this distribution of tasks is the specialization of labor. Adam Smith, the father of economics, explained this in his seminal work, The Wealth of Nations. Smith’s insight, demonstrated in his famous pin factory example, was that dividing production into discrete steps transformed output not incrementally but by orders of magnitude. Ten workers, each responsible for a single operation in pin manufacturing, could produce 48,000 pins a day. One worker attempting the full process alone might produce twenty.
The gains do not come from working harder. They come from working narrower. Each person developed precision in one step, eliminated the time lost switching between tasks, and eventually noticed improvements that a generalist spread across the full process would never have the vantage point to see.
It’s a simple concept, with profound consequences. Specialization of labor inherently invites complexity when you start to view systems at scale.
Systems don’t start complicated. They accumulate complexity because each generation of problems generates a new layer of solutions, and each solution creates new constraints for others to navigate.
The container ship that reduced shipping costs required deeper ports. The deeper ports required new cranes. The new cranes required new labor arrangements. The labor arrangements required new safety regulations. The safety regulations required enforcement mechanisms. Each link in the chain was a reasonable response to the problem immediately before it. Nobody planned the final result. The complexity was the residue.
This pattern repeats across every domain where modern systems have developed. Banking regulations respond to financial crises, which themselves emerged from earlier financial innovation operating without adequate oversight. Environmental standards respond to industrial harms that emerged when earlier industrial expansion had none. Healthcare protocols accumulate layer by layer as each set of interventions reveals failure modes the previous layer didn’t account for.
From the outside, looking at the full accumulated structure, it looks like purposeless sprawl. From inside, at any given moment, each layer was a specific answer to a specific problem. The person adding it was not thinking about the illegibility it would eventually contribute. They were solving what was in front of them.
This is why simplification is harder than it sounds. What looks like unnecessary complication from one vantage point is often load-bearing structure from another. A regulation that seems arbitrary to an individual consumer may be resolving a coordination problem between two industries that never interact with consumers directly. A financial instrument that seems needlessly complex may be distributing risk in a way that allows lending to occur at lower rates for people who benefit from it without knowing why. Remove a layer and you may streamline one interaction while breaking three others that were invisible to whoever proposed the simplification.
Some regulatory layers are genuinely vestigial: maintained by inertia or incumbent interest rather than function. Some complexity does serve those inside the system at the expense of those outside it. That is real, and identifying it matters. But the starting point for useful analysis is not the assumption that complexity is pure waste. It is the question of what each layer is actually doing.
The asymmetric ledger
There is a consistent pattern in how people evaluate complex systems. When a system produces a confusing or harmful outcome, the system gets blamed. When it produces a good outcome, the outcome is simply expected.
This asymmetry runs in one direction, reliably, and it is not a coincidence.
The same supply chain that frustrates you when a delivery is late also produced the product you ordered at a price that would have been unthinkable to previous generations. The same labor market that makes job searching feel arbitrary and opaque coordinates millions of employment matches every year without a central planner directing any of them. The same tax structure that nobody can fully explain funds the roads, emergency services, courts, and public health infrastructure that most people would not voluntarily surrender.
But the asymmetry runs deeper than credit and blame. It operates through a visibility problem that is structural.
Harms are visible and local. Benefits are invisible and distributed.
When a supply chain delivers a product late, one person experiences the problem directly, at a specific moment, with a clear emotional charge. When the same supply chain delivers ten thousand products on time at competitive prices, the outcome disperses across ten thousand transactions, each individually unremarkable, none of them experienced as a gift from the system. Harm concentrates, but benefit disappears into the void.
You do not feel the child mortality statistic that improved across a generation. You feel the delayed package. That gap is not a failure of rationality. It is a structural feature of how humans encounter systems larger than any individual’s view of them. The distributed nature of the benefits makes them functionally invisible to direct experience. You would have to construct a mental model of the counterfactual to feel the benefit: imagining the shirt arriving not for twelve dollars but for sixty, or not arriving at all. Most people don’t do that, not out of intellectual laziness, but because life doesn’t naturally prompt the comparison.
Attribution compounds this. When a business produces an affordable product, we credit the entrepreneur’s vision, the brand’s quality, the market working as it should. When the same business produces a confusing outcome for customers, we credit the system’s malice or incompetence. The infrastructure that made the entrepreneur’s product viable in the first place becomes wallpaper. The same pattern plays out in public life: government services that work fade into the background; government processes that burden people are visible and memorable. Both experiences are real. Only one of them shows up in the running tally.
The result is an evaluation that is structurally skewed toward the costs of complex systems and away from their outputs. A system producing substantial value alongside real friction will tend to be experienced primarily through the friction. This is not a defense of the friction or an argument that concentrated harms are acceptable. It is the observation that if you want to evaluate a system accurately, you have to deliberately ask what it is producing that isn’t showing up in your immediate experience of it.
Load-bearing or vestigial
The question “why is this so complicated?” is almost always the wrong entry point. It treats complexity as a design failure rather than as the residue of a trade-off. Trade-offs can be evaluated. Design failures can only be condemned.
The question that actually opens something up is what the complexity is paying for.
Sometimes the answer is: not much. Some regulatory layers exist because they were never dismantled after the problem they were designed to solve disappeared. Some intermediary structures persist because the incumbents who benefit from them have more leverage over the system than the people bearing the costs. Trace the function of each component and genuine waste becomes visible. These are real findings, and they matter.
But often the answer is: more than you’d expect. The opacity that frustrates you may be the same structure enabling coordination between parties who would otherwise have no way to transact. The abstraction that loses you may be what makes a global product affordable to someone who couldn’t otherwise access it. The process that feels pointlessly burdensome may be preventing a category of failure you have never experienced precisely because the process is preventing it.
The shirt arrives at your door for twelve dollars. The system that made it possible also frustrates you daily. Those two things are inseparable. The interesting question is not whether to tolerate the complexity in some general sense. It is more specific: which layers are load-bearing, for whom, and what would actually disappear if they were removed?
So what?
Complexity is not the problem. Misread complexity is. Systems can be improved, simplified, and made more accountable; but only when the analysis starts from what each layer is actually doing rather than from the assumption that complication is waste. That distinction sounds small. Its practical consequences are not.
Simplification is possible and indeed desirable in almost every circumstance; that doesn’t mean it is simple.
A first-principles approach to understanding a problem is the solution. First principles break things down to their basic tenets. The idea comes from physics and engineering: strip away assumptions, inherited conventions, and accumulated complexity until you reach what is actually and irreducibly true. The properties of the materials. The constraints of the problem, not the constraints of the existing solution. From those foundations, you rebuild. The advantage is that you are no longer constrained by the architecture of what already exists. You can see which layers of the current structure are load-bearing and which are artifacts of earlier decisions that stopped serving their purpose long ago. Even if that means breaking things down to the atomic level.
The challenge with applying first principles to human behavior and complex system interactions is that the atoms have individual personalities and incentives, often hidden, that need to be uncovered. This is cognitively hard work. Implementing changes, especially to established human behavior, is even harder.
In human systems, the first principles are incentives. Strip away the accumulated architecture of any complex system and you will find, underneath each layer, a set of actors responding rationally to the constraints in front of them. The complexity is not malicious. It is the aggregate output of individually sensible decisions, made by people with information you don’t have and objectives that were never aligned with yours. Complexity is not inevitable; but it is almost always the natural residue of incentive structures that were never redesigned, only extended.
This also determines who pays for the complexity. The friction you experience and the function it enables rarely fall on the same person. Someone added the layer because it solved a problem in front of them. Someone else absorbed the cost.
Most critiques of complex systems are, underneath the argument, really about who bears the friction. Few are about whether the friction is worth the outcome it produces. Those are different questions. The second one is much harder to orient to.
And only the second one is actually about the system.