Please follow us on Gab, Minds, Telegram, Rumble, Truth Social, Gettr, Twitter, Youtube
There is a certain kind of consultant who has existed in every era of American business. He arrives from outside, speaks with complete authority about an industry he has never worked in, leaves a presentation behind, collects a substantial fee, and is gone before anyone can establish whether the advice worked. In the 1990s he was selling offshoring. In the 2000s financial engineering. In the 2010s DEI transformation. The pitch never really changed — your current approach is inefficient, your competitors are already moving, the window to act is closing — only the product did. And the consequences, reliably, arrived after he was already in the next boardroom.
Today McKinsey, BCG, and Bain are selling AI-first. Silicon Valley is buying, and once Silicon Valley buys, every other corporate boardroom in the country follows. That is how the herd works.
The presentation has a specific shape. Generative AI can write production code and your competitors are already deploying it. You are paying two hundred engineers to do what a well-prompted model handles overnight. A new policy follows almost automatically: before any manager can justify a new hire, they must first demonstrate that AI cannot solve the problem. Engineering stops being the answer and becomes the last resort.
The promise underneath is that complexity disappears. You no longer need people who understand systems deeply, just a product manager with a good prompt. In that framing a large engineering team starts to look less like an asset and more like a habit nobody has questioned.
What nobody puts in the presentation is where the complexity goes. It doesn’t disappear. It sinks. Into layers of code no one feels ownership over and no one has time to question. The slide deck calls that efficiency. The postmortem will call it something else.
The railroad barons of the 1880s ran the most powerful private infrastructure in the country and managed their workforce the same way corporate America is managing its engineers now. A cost to be minimized rather than knowledge to be kept. When the political backlash came and Congress created the Interstate Commerce Commission in 1887, the people who staffed it and shaped how it operated were largely former railroad men. They knew the pricing structures, the routing decisions, the places where the whole system was held together by assumptions nobody had tested. The industry spent years pushing out the people who understood it from the inside, and those people did not forget what they knew on their way out the door.
A century later the Pentagon found itself with a version of the same problem. The Cold War ended, the threat assessment changed, and Washington decided it was carrying more engineering talent than the new mission required. Between 1987 and 1997 the defense industry shed roughly a million jobs. The engineers who left were not entry level. They had spent careers inside the most demanding technical programs ever built, in environments where failure was not a learning opportunity. When the tech boom pulled them westward they brought that seriousness with them, and it quietly shaped the engineering culture of the early commercial Internet. The Pentagon did not intend to seed its future competitor. It just decided, at a particular moment, that deep technical expertise was overhead it could no longer justify.
Silicon Valley was built in part on talent the government decided it no longer needed.
The engineers Silicon Valley is pushing out today are not bootcamp graduates. Many of them designed the distributed systems that move billions of dollars, built the cloud architectures that hospitals and governments now depend on, spent years learning not just how these systems are supposed to work but how they actually behave when something goes wrong at three in the morning. They know which parts of the architecture are solid and which were decisions made under deadline pressure five years ago and never revisited. That kind of knowledge does not live in documentation, it lives in the person.
These people are not going to sit at home. Defense contractors are in the middle of serious efforts to modernize software infrastructure that has not meaningfully changed since the 1990s. Cybersecurity firms are expanding against a threat landscape that keeps growing. Autonomous systems companies need engineers who understand production scale. Intelligence organizations want people who know how large platforms are actually built, because those platforms are now part of how geopolitical competition works. None of these employers are debating whether engineers are worth the headcount. They are recruiting aggressively for exactly the people corporate America just told were no longer necessary.
Someone will mention NDAs here. The assumption is that legal frameworks create meaningful walls between industries, that what an engineer knows about one company’s infrastructure stays there by contract.
A senior engineer leaving does not carry files. He or she carries an understanding of how the system actually works versus how it is documented, where the architecture holds up and where it does not, how data truly flows through the organization, and what the real exposure looks like under conditions nobody planned for. None of that requires violating an NDA, it is just an accumulated experience. Experience travels with the person who has it, into whatever room values it next.
You cannot litigate memory, you cannot retrieve intuition once it walks out the door.
For a long time the equilibrium worked in Silicon Valley’s favor. The best engineers preferred the private sector. The pay was better, the problems were interesting, and the culture rewarded people who understood systems deeply. The defense and intelligence communities could recruit from that pool, but they competed from a disadvantage.
That advantage is being given away. Not by a geopolitical rival, not by regulation, but by corporate consensus. If the private sector signals consistently that engineers are overhead, and the national security establishment signals consistently that engineers are essential, talented people will draw their own conclusions. They will not do it out of ideology or grievance. They will do it because that is what talented people do when one employer stops valuing what they know and another one starts calling.
Over time, the center of gravity shifts. The deepest operational knowledge of civilian infrastructure begins to concentrate in institutions whose incentives are strategic rather than commercial. That matters because those institutions shape procurement rules, influence cybersecurity standards, draft regulatory language, and define what counts as “national security” when pressure rises. When the people advising those decisions once built the platforms now being governed, the balance changes quietly.
Influence does not follow ownership. It follows understanding.
If Silicon Valley stops being the place where the system is understood best, it does not remain in control of how that system evolves. It becomes the terrain. And the people who once maintained it become the people mapping it from the outside.
The companies cutting engineers most aggressively are the ones with the most to lose from what those engineers build next.
A senior engineer who spent a decade inside a major cloud platform carries a detailed picture of how that infrastructure actually works, where it is resilient and where it is brittle, what assumptions were made under pressure, and what the real exposure looks like under stress. When that person joins a defense contractor or a government program, they bring something the law was never designed to contain. They bring an internal map of the terrain.
The defense establishment reduced its engineering base in the 1990s and watched much of that talent shape the commercial internet era. It did not collapse, it simply lost the position of being the place where the most ambitious engineers wanted to be. Influence followed talent.
Corporate America is now making its own version of that bet, that the expertise it sheds will not meaningfully alter the balance later. That the people who understand its systems best will remain aligned with its interests even after being told they are replaceable.
History suggests something else. When technical knowledge migrates, power migrates with it. Not overnight, not dramatically, but decisively.
McKinsey will not be around to answer for this one. They are already in the next boardroom, same charts, new product. The engineers will land on their feet, they always do.
The question is not whether they will succeed. The question is who will be setting the rules when they do.
Get the latest on NYC and the Empire State delivered straight to your inbox
