Edition
By Chris Blask
We keep mistaking concentrated dependence for prosperity. Then we act surprised when something breaks.
On Friday, the FAA grounded flights across the United States. A telecommunications disruption had interfered with the communications and radar information on which air traffic control depends.
The next day's official explanation was astonishing. During maintenance at a remote FAA site, a contractor was moving an old copper connection to fiber. Something in the change appears to have caused equipment to flood other routers with connection attempts. The precise fault—equipment, installation, or something else—was still under investigation.
Good. Investigate it. Find out exactly what happened. Hold the right people accountable.
But don't lose the larger question in the search for a bad router or a bad technician.
What kind of architecture allows an ordinary change at one remote location to threaten the operation of an entire national airspace?
The controllers and pilots did the right thing by putting safety first. The grounding was a necessary response to uncertainty. The disgrace, if the preliminary account holds, is a network whose failure could propagate that far in the first place.
The copper wire was never the problem. The dependency was.
In the same week, Jensen Huang stepped down after 33 years leading Nvidia. Colette Kress became interim CEO. There is no reason to assume she and the rest of Nvidia's people cannot carry the company forward. They deserve to be judged by what happens next.
But the transition exposes another question we have spent decades avoiding. If a new AI data center depends on one extraordinary executive personally deciding to back it, how much of that center's financing is actually a financing plan—and how much is a bet on the continuing availability and judgment of a single human being?
We have confused successful heroes with successful systems.
A bridge is not a foundation
Imagine connecting two continents with a single T1 line. Then build enormous cities, trading networks, hospitals, factories, and financial systems on either side. Commerce booms. Buildings rise. Charts go up and to the right.
Look at how prosperous we are! The continents are connected!
Except they're connected by one line.
That is not resilience. That's an outage patiently waiting for a date.
The same design mistake appears everywhere, dressed in different clothes. A software bill of materials full of libraries nobody understands or knows how to replace. A DNS configuration whose apparent redundancy disappears when one provider fails. BGP routes that are distributed in principle but hostage to a narrow set of operational dependencies in practice. A payment network with no usable alternative clearing path. A city that cannot function when one cloud region is unavailable. An AI application whose memory, tools, and identity evaporate when a vendor changes its terms.
Or an institution that runs beautifully, as long as the founder keeps getting out of bed and making all the right decisions.
We call these things efficient.
Often, they are merely efficient at hiding risk.
A system can be cheap to operate precisely because somebody else will absorb the cost when it fails: customers, travelers, employees, communities, taxpayers, or the next generation. We count the savings from centralization every quarter and leave the cascading failures off the balance sheet until they arrive.
Then we call the failure unprecedented.
No. The failure was built in.
A thousand tiny monarchies
The modern world has become oddly comfortable with indispensable people and institutions.
A few CEOs whose personal choices move enormous markets. A few hyperscalers whose services have become assumptions embedded in everybody else's services. A few financial intermediaries, chip suppliers, identity providers, and platform owners whose decisions reach into businesses they have never met and communities they cannot name.
Sometimes we call them visionaries. Sometimes oligarchs. Sometimes simply the market.
The label matters less than the dependency.
We keep building prosperity on the premise that these particular people and organizations will remain competent, available, benevolent enough, and willing to keep doing the same things indefinitely.
And when someone points out the danger, we get an absurd argument in return: without these concentrations of power, innovation would stop, growth would collapse, and civilization would be reduced to committees shuffling paper.
As though our only options were robber barons, kings, czars, and glorious corporate saviors on one side, or the gray halls of a central committee on the other.
What a poverty of imagination.
There are many ways to organize invention, investment, production, trade, and public life. Federations. Cooperatives. Competing firms. Shared infrastructure with open interfaces. Local ownership joined by interoperable networks. Private enterprise with real exit rights and independently enforceable obligations. Institutions that distribute decisions without dispersing responsibility into the fog.
None is magical. All require engineering and governance. But the choice between a monarch and a committee is not the menu of human possibility.
Scaffolding is not a city
Venice rests on timber piles driven into the lagoon centuries ago. Under the right waterlogged, low-oxygen conditions, wood can endure remarkably well. Those piles were not random props left to rot in the air; their durability depended on the material, the environment, and the structure working together. The city has also required continuous maintenance.
That is engineering.
Contrast it with a city cantilevered over a valley on wooden scaffolding that nobody has inspected since the first builders departed. People move in. Property values rise. A mayor announces a golden age. Investors discover that extending another platform is cheaper than examining the posts.
Soon everyone is too committed to ask what is holding the whole thing up.
We do this with software. We do it with capital. We do it with government. We do it with leadership.
Scaffolding is useful. A charismatic founder can be useful. A dominant supplier can be useful. A single cloud service can be useful.
The mistake is to leave the scaffolding in place and call it a foundation.
Biological systems are not perfectly redundant or immune to collapse. They have critical organs and vulnerable pathways too. But survival over time selects for adaptation, repair, substitution, and the ability to tolerate changing conditions. A flourishing ecosystem does not depend on one particular animal continuing to make every decision forever.
Our institutions should learn something from that.
Make the dependencies visible—and replaceable
The answer is not to ban big companies, abolish leadership, or insist that every village fabricate its own GPUs.
Scale has advantages. Specialization has advantages. Extraordinary people do extraordinary things.
The question is what rights and capabilities the rest of us retain when those advantages become dependencies.
Can you identify every critical software component and test what happens when it disappears? Can you move your data, identity, and institutional memory elsewhere? Can a community continue operating essential services when a provider goes down? Can the people affected by a system examine its consequential decisions? Can an organization change leaders without losing the knowledge, authority, and relationships that make it function?
Does redundancy actually follow independent paths, or do two supposed backups quietly share the same power supply, DNS operator, account, or permission system?
Can someone tell the CEO that the CEO is wrong—and can the organization correct course without treating that as an existential crisis?
Those are design questions. They are also economic questions, political questions, and questions about freedom.
A credible software bill of materials is not a decorative inventory. It is a tool for understanding and managing dependencies. Trustworthy AI memory is not merely an enormous archive; it must be portable, correctable, provenance-aware, and governed by people who can inspect what happens to it. A resilient network is not one whose central operator promises never to make a mistake; it is one that limits the damage a mistake can do.
Transparency matters here, but not as universal exposure of everyone's private information. We need visibility into the exercise of power and control over the disclosure of personal data. We need evidence that can be challenged, authority that can be bounded, and systems that can fail locally without failing everywhere.
That is not anti-business. It is a better business proposition.
Real resilience costs something to design. Fragility costs much more to live with.
The test of a founder
I am building a company. I know the temptation to think that if I keep doing what I do every day, everything will be fine.
That is ridiculous.
I might be absent tomorrow. I might be mistaken today. Someone else might have a better idea, or simply know something I don't.
If the company succeeds only while I continue performing the same daily miracle, I have not built a durable organization. I have built another single point of failure with a flattering biography.
The test is not whether the founder is indispensable at the beginning. Often the founder is.
The test is whether the company becomes less dependent on the founder as it grows: whether it preserves its knowledge, distributes legitimate authority, tests its assumptions, makes room for challenge, and can carry its purpose beyond any one person.
The same test applies to Nvidia, the FAA's network, our banks, our clouds, our governments, our AI systems, and the little organizations we are building in garages and offices everywhere.
Prosperity that can exist only while a handful of people and components remain exactly where they are is not prosperity secured. It is prosperity borrowed against a failure we haven't priced.
We can build differently.
We can still have invention, wealth, trade, ambition, and extraordinary achievement. We can have better systems because we stop treating fragility as the unavoidable price of all those things.
If replacing a piece of copper can threaten a continent's flights, the problem is not the copper.
If replacing a leader can threaten a civilization's confidence, the problem is not the leader.
The problem is what we decided to build around them.
And we can decide differently.
Canonical first-party Edition at QuietWire. Source artifact dated 16 November 2026. Canonical URL: https://www.quietwire.ai/editions/the-copper-wire-was-never-the-problem/.