PART I
The Waste
Billy Joel - Allentown
In 1942, the United States produced 48,000 military aircraft. Not over a decade. In a single year. Automobile factories that had been stamping out Chevrolets in January were building B-29 bombers by April. Appliance manufacturers retooled for ammunition. Shipyards that had never built anything larger than a fishing trawler were launching aircraft carriers, eight a month by 1944. A warplane rolled off the line every five minutes. 150 tons of steel every minute. Fifty merchant ships a day.
None of this required some unique genius that existed in 1942 and vanished afterward. The engineers were ordinary. The workers were ordinary. The companies, Ford, Chrysler, DuPont, Kaiser, were ordinary American firms doing ordinary American work. What was extraordinary was the direction. Someone pointed the full weight of American industrial capacity at a single objective, removed the obstacles between the workforce and the work, and let the country do what it does when the target is clear: build at a speed and scale that rewrites what everyone else considers possible.
That capacity is not gone. Americans did not become less capable between 1945 and now. There are more engineers today, better trained, with better tools, more computational power, more raw material to work with. The industrial base has shifted, we build software and systems now where we once built steel, but the underlying capacity to direct massive amounts of talent and capital at hard problems is not a historical artifact. It is latent. Dormant. Still here.
It has no target.
"Well we're living here in Allentown, and they're closing all the factories down." - Billy Joel, 1982
Or rather, it has thousands of targets, almost all of them trivial.
Let's start with AI, since it's the obvious one. The technology itself is extraordinary, genuinely the most significant tool since electricity, maybe since the printing press. For the first time in history, the marginal cost of cognitive labor approaches zero. A system that can reason, analyze, write, and synthesize across the entire body of recorded human knowledge is available to essentially everyone. This should be restructuring the economy from the ground up.
What is it actually being used for? Chatbots. Content generators. SaaS copilots that summarize your meetings and draft your emails. An endless cascade of thin wrappers around the same foundation models, each one raising venture money on the premise that their particular wrapper, the one for legal, the one for sales, the one for HR, will be the one that captures the market. Seventeen different AI tools for email productivity. Not one serious effort to apply the technology to, say, rebuilding the power grid.
There is a version of this technology that redesigns American manufacturing. That gives every small fabrication shop the production intelligence currently locked inside Siemens and GE. That automates the engineering analysis on infrastructure projects so that a bridge replacement doesn't take eleven years of environmental review before a single beam is laid. That takes the operational knowledge trapped in the heads of retiring master electricians and welders and encodes it into systems so it doesn't vanish when they walk out the door.
That version of AI is almost entirely unfunded. It doesn't fit the SaaS margin profile. Physical world problems are messy, they involve permits, concrete, regulations, legacy systems, and human beings who don't behave like enterprise software customers. The returns are slower. The TAM slides don't look as clean.
So the capital goes elsewhere.
In 2023, American venture capital deployed roughly $170 billion.
The bulk of it went where it always goes: software. SaaS, fintech, healthtech, which mostly means health software, not health infrastructure, adtech, e-commerce enablement, crypto. The categories that scale fastest, that have the highest margins, that produce the cleanest returns on the timeline a venture fund requires.
Energy infrastructure. Manufacturing technology. Construction. Water systems. Grid modernization. Agricultural technology. Defense. The categories that determine whether 330 million people can power their homes, drink clean water, eat, commute, and defend themselves, these collectively received a fraction of what went into software companies building tools for other software companies.
Nobody planned this. No one sat in a room and decided to systematically defund the physical economy. The incentive structure did it automatically. Venture capital optimizes for returns within a fund's lifecycle. Software compounds faster than steel. A SaaS company can reach $100M ARR without touching the physical world. A company trying to modernize water infrastructure has to deal with municipal procurement, federal regulations, and construction timelines measured in decades.
So the money flows to the path of least resistance, and the path of least resistance runs as far from the physical world as it can get, and the physical world, the pipes, the wires, the roads, the bridges, the grid, degrades quietly in the background.
The United States currently has over 46,000 bridges rated as structurally deficient. The average age of the American power grid is 40 years. The American Society of Civil Engineers has graded the country's infrastructure a C-minus, which is generous. Jackson, Mississippi, has experienced three major water system failures in four years. The average time to approve and build a highway interchange is approaching a decade.
These are not problems of capability. They are not even problems of money, $170 billion a year in venture capital alone, trillions more in institutional capital looking for returns. They are problems of direction. The capital exists. It is pointed at the wrong things.
The same misallocation operates on people, not just capital.
The surge in youth entrepreneurship over the last five years is massive, and on the surface it looks like the healthiest signal in American culture. More young people are starting businesses than at any point in recorded history. The instinct is right, reject the credential treadmill, bet on yourself, build something. That instinct is genuinely good.
But look at what they're building...
Social media marketing agencies. Thousands of them, in every city, offering the same service, run Facebook ads for local businesses, learned from the same YouTube guru, using the same playbook, targeting the same market. They're not independently arriving at the same business idea through parallel genius. The recommendation algorithm showed them the same video, which showed them the same course, which gave them the same template. Each one believes they made an independent decision. They did not.
This extends beyond agencies. Dropshipping stores. Print-on-demand brands. Coaching businesses that coach people on starting coaching businesses. "Faceless" social media accounts that post AI-generated content to sell digital products about posting AI-generated content. An entire economy of people selling the dream of the economy they're participating in, with nobody at the bottom producing anything.
There is a kid somewhere in America right now, there are thousands of them, with genuine technical capability, real work ethic, and the risk tolerance to build something hard. He is spending all of that building a personal brand about building personal brands. She is launching an agency that will be indistinguishable from the nine hundred agencies launched the same month. Not because they lack talent. Because the information environment that shaped their ambition, the algorithms, the incentive structures, the visible success stories, pointed them at the things that are easiest to start and fastest to monetize, which are also the things that contribute the least to the economy and the civilization.
This is not their fault. You cannot blame a generation for responding rationally to the incentive landscape they were handed. But you can name the waste for what it is. An enormous amount of entrepreneurial energy, the kind of energy that built Standard Oil, that built Ford, that built the semiconductor industry, is being absorbed by businesses that do not matter. Not because the people don't matter. Because nobody showed them what does.
The pattern here is rather more simple than would be expected.
America does not have a talent shortage. It does not have a capital shortage. It does not have a technology shortage. It has a direction shortage. Every resource that would be required to rebuild the country's industrial base, modernize its infrastructure, achieve energy independence, reshore critical manufacturing, and maintain military superiority, all of it exists, right now, in abundance.
The country that built the interstate highway system in a decade now takes seven years to approve a single interchange. Not because the engineers disappeared. Because the permitting system, the environmental review process, the litigation exposure, and the bureaucratic accumulation of sixty years of procedural complexity have made it structurally impossible to build at the speed the country once considered normal. The capability is there. It is buried under process.
The country that went from zero to nuclear weapons in three years, a problem that had never been solved by any civilization in the history of the world, now takes fifteen years to produce a fighter jet that still doesn't work properly, at a cost that exceeds the GDP of most nations. The F-35 program is not a failure of engineering. The engineers are as good as they've ever been. It is a failure of the system that manages the engineers, a procurement apparatus so encrusted with process, compliance, and institutional self-preservation that the actual objective (build the plane) has become secondary to the process of building the plane.
In 1945, the Manhattan Project employed about 125,000 people and produced the most consequential weapon in human history for roughly $2 billion, about $30 billion in today's dollars. The F-35 program has consumed an excess of $2 trillion across its lifetime. One of these produced a technology that ended a war and reshaped the global order. The other produced a plane that can't reliably fly in thunderstorms.
The comparison is not entirely fair, they are different programs solving different problems in different eras. But the scale of the disproportion tells you something about what happens when a system optimizes for process instead of output. The process becomes the product. The meetings become the work. The compliance becomes the mission. And somewhere underneath all of it, the actual thing that was supposed to get built waits.
What makes this different from ordinary decline, the kind that happens to every civilization eventually, is that it's happening at the precise moment when the tools to reverse it are better than they have ever been.
AI could restructure how America builds things. Modular nuclear could solve the energy problem within a decade. Advanced manufacturing techniques, 3D printing, robotic fabrication, AI-driven quality control, could reshore production that was offshored a generation ago. The technologies are not theoretical. They exist. Some of them are mature. The obstacle to deploying them is not technical. It is institutional, regulatory, and directional.
Which means the waste is a choice. It is not a natural law. It is not entropy doing what entropy does. It is the accumulated result of thousands of small decisions, about where to allocate capital, about what to reward, about what to permit and what to obstruct, about what to teach and what to ignore, that collectively produced a system which is extraordinarily good at generating software and extraordinarily bad at building physical things.
The country is not broken. The country is misdirected. Every piece of evidence that looks like decline, the crumbling bridges, the impossible timelines, the talent building trivial businesses, the capital chasing trivial returns, is actually evidence of misallocation. The resources are there. They are pointed at the wrong targets by systems and incentives that nobody designed on purpose but that nobody is fixing either.
This distinction matters because it determines whether the right response is despair or redirection. If America were actually declining, running out of talent, running out of capital, running out of energy, then the prognosis would be terminal. But that is not what is happening. What is happening is that a country with more capability than any civilization in history is wasting that capability on things that do not matter, while the things that do matter, energy, infrastructure, manufacturing, defense, the physical backbone of a functioning nation, are starved of attention, capital, and talent.
That's the waste. Not a shortage. A misdirection. And misdirection, unlike decline, can be corrected.
The question is who corrects it, because the people and institutions currently directing the country's energy are the same ones who misdirected it in the first place. Which brings us to the question of who, exactly, is in charge.