Last week I showed you the NUMMI proof — the same workforce, in the same building, going from GM’s worst plant to GM’s best in a single year. The only variable that changed was the operating system. The intelligence was always there. The system had suppressed it.

Some readers pushed back. They accepted the NUMMI data but questioned the word “suppression.” That implies intent, they said. Nobody sat down and designed a system to make workers dumber. What happened at Fremont was just bad management. Entropy. Dysfunction. Not a deliberate architecture of intelligence removal.

This week I’m going to show you that it was, in fact, deliberate. That the dominant Western manufacturing operating model was designed — on paper, in published monographs, with explicit instructions — to remove discretion, judgment, and initiative from the people doing the work. That the architect of this system said so, in terms so plain they barely require interpretation.

His name was Frederick Winslow Taylor. His 1911 book, The Principles of Scientific Management, is the most influential management text of the twentieth century. And it is, if you read it honestly, a confession.

The Problem Taylor Was Solving

To understand what Taylor built, you have to understand what he saw.

Born into a wealthy Philadelphia family in 1856, Taylor took an apprenticeship in a foundry during the Panic of 1873 and worked his way up through the Midvale Steel Company. He was a mechanical engineer by training, an obsessive by temperament, and a man who believed — with the fervor of a convert — that inefficiency was a moral failing.

What bothered Taylor most was what he called “soldiering” — the deliberate restriction of output by workers. He saw two varieties. “Natural soldiering” was the human tendency to take it easy. “Systematic soldiering” was the organized, shop-wide practice of controlling output to protect jobs and prevent management from raising quotas.

Taylor watched skilled workers hold back. He watched them teach new workers to hold back. He concluded that this was a fundamental problem of information asymmetry: the workers knew how the work was actually done, and management didn’t. As long as the knowledge of the craft lived inside the workers’ heads, management could never truly control production.

His solution was breathtaking in its ambition and chilling in its implications: transfer all knowledge of the work process from the workers to management.

The Architecture of Suppression

Taylor didn’t hide his intent. He published it.

The core mechanism of scientific management was the separation of planning from execution. Managers would study every task, break it into its smallest components, determine the single best method for each component through time-and-motion analysis, and encode that method on instruction cards. Workers would then follow these cards exactly. No deviation. No improvisation. No judgment.

In a 1906 lecture, Taylor was asked about worker initiative under his system. His answer has been preserved in multiple sources, and it bears repeating in full because of what it reveals about the philosophical foundation of the operating model that still governs most factories in the Western world:

“In our scheme, we do not ask for the initiative of our men. We do not want any initiative. All we want of them is to obey the orders we give them, do what we say, and do it quick.”

This was not a slip. It was the design specification.

Taylor’s method required that every job be decomposed so thoroughly that “no one worker would possess any knowledge which might be unique enough to put this worker in a position of power vis-à-vis management.” The explicit goal was to eliminate the worker’s informational advantage — to strip away the craft knowledge that gave labor any leverage in its relationship with capital.

The instruction cards were the mechanism. Before Taylor, a skilled worker carried the accumulated knowledge of how to do the job — the speeds, feeds, angles, sequences, and judgment calls that made production work. Taylor’s time-and-motion studies extracted this knowledge, codified it, and handed it to a new class of planning-department clerks. The worker was left with a card that said what to do, how to do it, and how long it should take.

What had been a thinking person performing a complex task became a pair of hands following a script.

The Schmidt Parable

Taylor’s most famous — and most revealing — illustration of his philosophy involves a pig-iron handler at Bethlehem Steel whom he called “Schmidt” (the man’s real name was Henry Noll).

Schmidt’s job was to load pig-iron ingots onto railcars. Taylor studied the task, redesigned the work method, and claimed to have increased Schmidt’s daily output from 12.5 tons to 47.5 tons. This was Taylor’s showcase proof of concept — the story he told in congressional testimony and in The Principles of Scientific Management to demonstrate the transformative power of his system.

But listen to how Taylor described the kind of worker he needed for this job. He was looking for a man, he wrote, “so stupid and so phlegmatic that he more nearly resembles in his mental make-up the ox than any other type.” Taylor specified that the ideal pig-iron handler should be “merely” strong and obedient — someone who “shall be so stupid that the word ‘percentage’ has no meaning to him.”

Taylor wasn’t describing a finding. He was describing a requirement. The system he built needed workers who wouldn’t think, because thinking workers would resist the removal of their discretion. The architecture demanded compliance, and Taylor was honest enough to say so.

What he didn’t acknowledge — what the next century of evidence would prove — was that the “stupidity” he sought was not a characteristic of the workers. It was an output of the system he imposed on them.

The Deskilling Machine

Taylor’s system didn’t just ignore worker intelligence. It actively dismantled it.

Before scientific management, factory work was organized around craft. A skilled machinist understood the properties of different metals, the behavior of cutting tools at various speeds, the relationship between feeds and finishes. This knowledge lived in the worker’s hands and head, accumulated over years of apprenticeship and practice. It gave the worker autonomy, bargaining power, and — critically — a sense of ownership over the work.

Taylor saw all of this as a problem to be solved.

His method systematically transferred craft knowledge from the shop floor to the planning office. He created a new division of labor: managers think, workers do. The foreman’s traditional role was broken into specialized functions — speed boss, quality inspector, instruction-card clerk, shop disciplinarian — all reporting to the planning department. Workers were left with tasks so simplified that, in Taylor’s own framing, almost anyone could perform them.

The sociological literature calls this “deskilling.” The historian Harry Braverman, in his landmark 1974 work Labor and Monopoly Capital, argued that Taylorism represented the deliberate degradation of work — a transfer of knowledge and control from labor to capital that had nothing to do with efficiency and everything to do with power.

But you don’t need Braverman to see it. Taylor himself described the process with an engineer’s precision. He created instruction cards that “noted each motion and each decision that workers had passed along informally to each other.” He recorded these decisions for management’s use. He then assigned the simplified, decision-free residue of the job to the workers and called the result “scientific.”

What he had actually done was build a machine for converting intelligent workers into compliant ones.

The Operating System We Inherited

Here is what makes Taylor’s confession relevant 115 years after publication: the operating model he designed is still running.

Not in its pure form. Not with stopwatches and instruction cards and Schmidt loading pig iron. But in its deep structure — in the separation of planning from execution, in the assumption that managers think and workers do, in the reflexive removal of discretion from frontline jobs, in the belief that standardization means the elimination of judgment rather than its foundation.

Every time a plant manager says “just follow the standard work” without asking what the worker sees that the standard doesn’t capture, Taylor is in the room. Every time an engineer designs a process that assumes the operator will contribute nothing beyond physical motion, Taylor is in the room. Every time a company invests millions in automation to eliminate frontline jobs while spending nothing on deploying the intelligence those workers already possess, Taylor is in the room.

The language has changed. We say “lean” instead of “scientific management.” We say “standard operating procedure” instead of “instruction card.” We say “operator” instead of “hand.” But the underlying assumption — that the purpose of the system is to make worker judgment unnecessary — persists in factories, warehouses, distribution centers, and service operations across the developed world.

It persists because it is embedded not in any single manager’s philosophy but in the structure of the operating model itself: the org charts, the job descriptions, the training programs, the performance metrics, the capital allocation decisions. It’s the water the fish swim in. Most managers enforcing the Taylorist model don’t know they’re doing it, any more than a fish knows it’s wet.

The Inversion Toyota Performed

This is why NUMMI matters so much as a companion to the Taylor story.

Taylor designed a system that said: extract the intelligence from the workers, encode it in management processes, and reduce the worker to an executor of prescribed tasks.

Toyota designed a system that said the opposite: build intelligence into the workers, give them tools to solve problems at the point of occurrence, and treat management’s job as creating the conditions for frontline capability to flourish.

Toyota’s standardized work is not Taylor’s instruction card. In the Toyota system, the standard is a baseline — the current best-known method, documented so it can be improved. And the person who improves it is the worker doing the job. The standard isn’t a ceiling on worker judgment. It’s a floor.

That distinction — the standard as ceiling versus the standard as floor — is the difference between intelligence suppression and intelligence deployment. And it explains why the same workers at Fremont, operating under the same union contract, in the same building, produced opposite results under the two systems.

Under Taylor’s architecture, the rational worker response is disengagement. Why invest your intelligence in a system that doesn’t want it? Why solve problems you weren’t asked to solve, in a structure that will give you no credit and may punish you for deviating from the card?

Under Toyota’s architecture, the rational worker response is contribution. The system asks for your ideas. It implements them. It credits you. It trains you to find more. The andon cord — the ability of any worker to stop the entire production line — is not a quality tool. It is a philosophical statement: your judgment matters more than the schedule.

Why This Matters for What Comes Next

I’m building this series toward an argument about AI and the future of frontline work. But I can’t make that argument until the historical foundation is secure.

Here is what the historical foundation says:

The system that governs most frontline work in the Western world was designed to suppress worker intelligence. Not as a side effect. Not as an unintended consequence. As the core mechanism. Taylor published the blueprint. He named the philosophy. He specified that he did not want initiative from his workers and that the ideal worker would resemble an ox.

For 120 years, we have been operating variants of this system and interpreting the results as evidence that frontline workers lack capability. We watched them disengage and called it laziness. We watched them stop solving problems and called it low skill. We watched them resist and called it bad culture.

NUMMI proved that the same people, in a different system, will solve problems, generate ideas, and produce world-class quality. The Taylor history explains why the old system produced the opposite: it was designed to.

The skeptic who says “there is no suppressed intelligence on the factory floor” is not wrong about what they observe. They are wrong about what they conclude. The disengagement is real. The absence of initiative is real. But these are outputs of a system, not characteristics of the people inside it.

The intelligence is there. It has always been there. The blueprint just told us to ignore it.

Next week: “The Suggestion Gap” — What happens when you actually ask frontline workers to think. Toyota gets a million ideas a year. Most Western manufacturers get silence. The difference isn’t the workers.

Dr. Venki Padmanabhan is a plant manager with 36 years of global manufacturing leadership experience, including executive roles at GM, Chrysler, Mercedes-Benz, Royal Enfield, and Ather Energy. He holds a PhD in Industrial Engineering from the University of Pittsburgh. He is the author of the forthcoming Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away and co-founder of the Capability Capital Institute. He writes The Long Game at thelonggameforall.substack.com.

Fortune / Travis Kalanick, March 19, 2026

Travis Kalanick — the man who built Uber by reclassifying taxi drivers as independent contractors so he wouldn’t have to pay them benefits — went on a podcast this week and called human workers “valuable.” He said that until super AGI arrives, humans will be — his phrase — “the long pole in the tent to progress.”

The long pole in the tent. That’s a project management term. It means the bottleneck. The constraint. The thing everyone is waiting on and trying to engineer around.

He called my people a bottleneck. And he thinks it’s a compliment.

Yesterday I wrote about Larry Fink spending a hundred million dollars to train electricians because he knows you can’t build the future without them. Today I want to talk about a man who looks at the same workforce and sees a problem to be solved.

Kalanick’s example is plumbers. If every other job were automated, plumbers would become “extremely valuable” because buildings can’t go up without them. The plumber isn’t valued for her judgment or her knowledge of which contractors cut corners in 2014. She’s valued because a robot hasn’t learned to sweat copper in a crawl space yet. The moment it does, she joins the pile.

This is a stay of execution called a white pill.

And when the tech prophets are pressed on what comes after, they reach for science fiction. Hassabis talks about colonizing the galaxy. Altman imagines spaceships and “super interesting” jobs in the solar system. This is not a workforce strategy. This is anesthesia.

But we don’t have to speculate about how Kalanick treats frontline workers. We have it on tape.

February 2017. Dashcam. Uber Black car, San Francisco. The driver is Fawzi Kamel — with Uber since 2011, one of the originals. And he has the CEO in his backseat.

Fawzi had watched for six years as Uber cut his fares. Each cut came out of his pocket, not Kalanick’s. Kalanick’s net worth went up with every reduction — lower prices, more riders, higher valuation. Fawzi’s income went down for the same reason. The math was elegant. The math was also extraction.

No grievance process. No union. No andon cord. Uber calls you a “partner,” but partners negotiate terms. You can’t negotiate anything.

So Fawzi speaks up. “I lost ninety-seven thousand dollars because of you. I’m bankrupt because of you.”

Kalanick’s response: “Some people don’t like to take responsibility for their own s**t. They blame everything in their life on somebody else. Good luck.”

Then he slammed the door. Fawzi gave him one star.

This is the same man whose company settled misclassification lawsuits — eight million in California, a hundred million in New Jersey, three hundred and twenty-eight million in New York. Who spent two hundred million on a ballot initiative to avoid calling drivers employees. Whose leaked documents showed him saying “violence guarantees success” about sending drivers into a dangerous protest.

This is the man now calling workers valuable.

When Fawzi told the CEO the system was destroying him, he was doing what any frontline worker does when the process is broken — he pulled the andon cord. Kalanick told him to take responsibility. And slammed the door.

I’ve met a lot of people Kalanick would call long poles. I wrote about one of them in an essay called “The Light That Was Already On.” Wally Vinton — worst team leader in my Trim Shop at Lansing Grand River. Management wanted him gone. Instead of writing him up, I walked onto the floor and Wally handed me the harness and said: “Try this.” Every problem he’d screamed about was real. So instead of slamming the door, we brought him support. Six months later, his area had the lowest downtime in the shop. Same person. Same paycheck. Someone listened.

Imagine if Kalanick had done that with Fawzi. “Show me. Show me what the fare cuts look like from your seat.” What would Uber look like if the company had spent six years investing in its drivers’ knowledge instead of six years classifying them out of legal protection? We’ll never know. Because he slammed the door.

After thirty-six years in manufacturing, here is what I know: labor is not a depreciating asset. It is an appreciating one — but only if you deploy it. Wally’s knowledge was compound interest. Every year on that line, he accumulated intelligence no manual captured — which paint booth drifted, which part numbers failed, where the injuries would surface next quarter. Fawzi’s knowledge was the same kind. Six years of every street in San Francisco. Patterns the algorithm couldn’t match. All of it compound interest. And the CEO told him to take responsibility.

Multiply Wally by every plant in America. Multiply Fawzi by every gig worker on every platform. That’s not the long pole in anybody’s tent. That’s the tent.

What Kalanick cannot see — because his career was built on extracting value from workers rather than investing in them — is that the opportunity is not replacing workers. It’s forming them. Not training, which implies a deficit. Formation, which implies a capacity. You don’t become a master by surviving automation. You become one by being invested in, tested, and trusted over time.

Yesterday, Larry Fink wrote a hundred-million-dollar check because he understands that the people who build things with their hands are not bottlenecks. They are the foundation. Today, Travis Kalanick called those same people the long pole in the tent.

The long pole is not a compliment, Travis. It means you’re the constraint everyone is trying to engineer away.

Wally was never the constraint. He was the foundation. And the people who can’t see that — the ones who slam doors and spend two hundred million to avoid calling workers employees and promise space colonies instead of decent wages — those are the ones who’ll spend the next decade wondering why their tent keeps falling down.

Dr. Venki Padmanabhan • The Long Game

Response to Fortune / BlackRock, March 2026

Larry Fink just wrote a hundred-million-dollar check to train electricians. The CEO of BlackRock — the man who manages eleven trillion dollars in assets — looked at the American economy and decided the most urgent investment he could make wasn’t in AI models or data centers or semiconductor fabs. It was in people who work with their hands.

That should stop you cold. When the largest asset manager on earth treats skilled tradespeople as the critical constraint on the future, something has shifted.

Fink told his Infrastructure Summit this week that the four-year college degree is cracking as a pathway to a stable career. That AI is going to disrupt the entry-level white-collar jobs that college was supposed to lead to. And that the real demand — the demand that isn’t going away — is for electricians, HVAC technicians, plumbers, and ironworkers. BlackRock committed a hundred million to train fifty thousand of them over five years.

He also warned that the class of 2026 could face the highest unemployment among new graduates in years — even without a recession. The jobs they trained for are being automated. The jobs that can’t be automated are the ones they were told not to pursue.

“The key for life for everyone is to find their purpose,” Fink said. “For some people, their purpose will remain to get a four-year or advanced degree — but that’s not going to be the pathway for everybody.”

The numbers are staggering. Electrical work accounts for up to seventy percent of data center construction costs. Over three hundred thousand new electricians are needed in the next decade just to meet AI-driven demand. At the same time, more than two hundred thousand electricians are expected to retire. Microsoft’s president has called the electrical talent shortage the single biggest challenge for data center expansion in the United States. Union journeyman electricians near Washington, D.C., earn over a hundred and twenty thousand a year plus benefits — and that’s before overtime.

So the machines that are supposed to replace human workers can’t actually get built without human workers. The AI revolution runs on electricity, and electricity runs on electricians. Try automating that.

But here’s where I part company with Fink, and it’s where my thirty-six years on the factory floor speak louder than his eleven trillion dollars.

A hundred million for training is better than nothing. It’s better than what most companies are doing, which is precisely nothing. But training and formation are not the same thing. Training fills a skill gap. Formation builds a human being. Training can be compressed into a bootcamp. Formation cannot be compressed any more than you can compress a pregnancy by adding resources. It takes the time it takes because the thing being built — judgment, pattern recognition, the ability to diagnose a system by listening to it — grows at human speed, not investor speed.

The German guild system understood this. An electrician apprenticeship in Germany takes three and a half years. Not because the curriculum is longer, but because the formation is deeper. You learn under a Meister — a master craftsperson who has been certified not just in the trade but in the teaching of the trade. You rotate through shops. You encounter problems you haven’t been trained for and learn to think your way through them. By the time you’re done, you don’t just have a skill. You have judgment. And judgment is the one thing that compounds.

America dismantled that system. We did it in three moves. First, the shareholder value revolution of the 1980s made quarterly returns incompatible with five-year apprenticeships — why invest in a worker who might leave? Second, the flexibility ideology turned institutional commitment into a liability. Third, the college-for-all movement stripped the trades of social prestige and funneled an entire generation into four-year degrees that many of them didn’t need and couldn’t afford. The result is exactly what Fink is now seeing: a country that needs three hundred thousand electricians and has no system to produce them.

I manage a plant in Wooster, Ohio. Sixty people, three production lines, corrugated stormwater pipe. My maintenance team is skeletal — one quasi-electrician and a couple of millwrights keeping aging equipment north of seventy-five percent OEE. Every piece of institutional knowledge that walks out the door takes years of compound intelligence with it. Every new hire starts from zero. There is no guild, no Meister, no formation pathway. There is only on-the-job training — which means trial and error, which means scrap, which means downtime, which means lost money.

Fink can see the macro problem. I live the micro version of it every day.

Labor is capital in formation. Capital is crystallized labor. That’s not a slogan — it’s an accounting identity that every foreman in America understands intuitively and no balance sheet captures. When an electrician spends five years learning to trace a fault by sound and smell and the way a breaker feels under her hand, that’s capital being formed. When she retires and nobody has been formed to replace her, that’s capital being destroyed. Fink’s hundred million is an attempt to rebuild what we spent forty years demolishing.

It’s a start. But a check to nonprofits is not a system. A system is what the Germans have — chambers, standards, masters, pathways. A system is what American manufacturing had in the 1950s when companies trained their own people and kept them for decades. A system is what we need to build again, not through nostalgia, but through architecture.

Tomorrow I’m going to tell you about a man who looked at the same workforce Larry Fink is trying to rebuild and called them a bottleneck. Travis Kalanick went on a podcast this week and called workers “the long pole in the tent.” Fink wrote a hundred-million-dollar check. Kalanick slammed a car door. Same week. Same workforce. Opposite conclusions.

The foreman already knows what Fink just learned. The electrician’s intelligence isn’t a cost to be minimized. It’s an asset that compounds. Every year on the job, she gets better — not because someone trained her, but because the work itself formed her. And when she leaves, the loss isn’t a line item. It’s a crater.

Larry Fink spent a hundred million dollars to learn what every foreman already knows: you can’t financialize your way out of a formation crisis. You have to build people. And building people takes time, trust, and a system that values what they become.

On Friday morning, the Bureau of Labor Statistics reported that the American economy shed 92,000 jobs in February. Markets had expected a gain of 50,000. The miss was 142,000 jobs. In a single data release, the question that has been building for months became impossible to ignore: Is AI destroying jobs faster than anyone admitted? Or is something else happening entirely?

I want to give you a precise answer. Not a reassuring one. A precise one.

Because the 92,000 number is not a single story. It is three stories running simultaneously — three separate mechanisms that produce the same headline but require entirely different responses. Conflating them is how the people responsible avoid accountability. Separating them is how the rest of us start to see clearly.

• • •

The first story is the one you have been sold.

Artificial intelligence has arrived. It is replacing human work at speed. Companies that once needed a hundred people now need ten. This is either the most exciting productivity revolution in history or an unfolding catastrophe for working people, depending on whom you ask — but everyone agrees it is real, it is happening now, and it is accelerating.

Except that is not what the data shows.

According to Challenger, Gray & Christmas, approximately 55,000 layoffs in all of 2025 were attributed directly to AI. That is less than one percent of all job separations for the year. A paper from the National Bureau of Economic Research found that ninety percent of executives — nine out of ten people actually making hiring and firing decisions — said AI has had no material impact on workplace employment over the past three years.

Brad Conger, chief investment officer at Hirtle Callaghan, manages twenty-five billion dollars on behalf of institutional clients. He is not a skeptic of AI. He has purchased five AI software products in the past six months. And he is blunt: “A job does 100 things in a day, and that’s a lot more than a single AI workflow can perform. It replaces activities that are just pieces of jobs.” At his firm, he says, not one position has been eliminated because of AI.

This is the first face: the Displacement Story. It is, for the most part, a fiction — or at minimum, a dramatic exaggeration of a real but limited phenomenon. AI is improving workflows at the margins. It is not yet dismantling workforces at scale. The companies claiming otherwise are, in Altman’s own word, washing.

But if AI is not actually doing the work of elimination, something is. Which brings us to the second story.

• • •

When Amazon announced 30,000 job cuts between October and January, its capital expenditures were on a trajectory from fifty-three billion dollars in 2023 to a projected two hundred billion in 2026. When Workday laid off 1,700 employees — eight and a half percent of its workforce — in February, CEO Carl Eschenbach declared the cuts were necessary to “prioritize AI investment and free up resources.” When Block laid off 4,000 people in early March, the company was simultaneously accelerating its investment in AI tooling.

Conger connects the dots directly: “AI’s not replacing jobs, but job cuts are funding AI expenditures.”

This is the second face: the Capex Trade. It is more honest than the Displacement Story, but only slightly — because it obscures what is actually being traded away. When a company eliminates experienced people to fund infrastructure bets, it is not making a neutral financial reallocation. It is liquidating a specific kind of asset — accumulated human knowledge, judgment, and institutional memory — to purchase a speculative one.

In manufacturing, we have watched this calculation fail for thirty years. A plant eliminates its senior machinists to cut costs. Five years later, nobody can diagnose the recurring vibration problem in line three, because the three people who understood it have been gone for years. The knowledge did not transfer. The savings were real. The capability gap that followed was also real, and it cost far more to close than the payroll savings ever generated.

The Capex Trade is executing that same logic at civilizational scale. Companies are liquidating the people who generate new knowledge to purchase systems that can only reproduce existing knowledge. The models improve at pattern-matching. They do not improve at the thing a forty-year machinist discovers in year forty-one — the insight that surprises even him, the knowledge that has never been written down anywhere and therefore cannot be trained on. When you eliminate the person, that knowledge does not go dormant. It disappears.

• • •

Then there is the third story. The one that neither the displacement narrative nor the capex analysis accounts for. The one that requires thirty-six years on factory floors to see clearly.

I have been in rooms where automation investments are approved. I have been on floors where the consequences land on actual people. And here is what I can tell you: the layoffs that get labeled AI-driven are disproportionately layoffs of people who were never given the chance to demonstrate what they were worth.

Aaron Zamost, Block’s own former head of communications, said it plainly after the March layoffs: the eliminated roles were disproportionately in policy and diversity functions — “standard prioritization and cost management,” not AI transformation. The AI story was cover. But cover for what, exactly?

Cover for this: most organizations have spent years treating their workforce as a cost to be minimized rather than an asset to be developed. Workers are trained to execute, not to think. Their ideas go unsolicited. Their judgment goes untrusted. Their intelligence — which compounds with every year of experience — is quietly suppressed by management systems designed for compliance rather than capability.

Then AI arrives. And suddenly the organization has a clean story for what is actually a very old failure. We are not eliminating people because we never bothered to develop them. We are eliminating people because technology has made them obsolete.

This is the third face: Leadership Laundering. The jobs being cut are disproportionately the jobs the organization never built. The workers being discarded are the workers the organization never invested in. The capability being outsourced to machines is the capability that was already paid for, in years of wages and hours of institutional time, and never deployed.

You cannot claim AI replaced your workers’ contribution if you never measured that contribution. You cannot say a machine made your people redundant if you never made your people essential.

• • •

I said I wanted to give you a precise answer, not a reassuring one. Here it is.

The 92,000-job number almost certainly reflects all three forces operating simultaneously, in proportions nobody can yet disentangle. Some of it is genuine AI-driven efficiency, narrow and real. Some of it is the Capex Trade — companies funding infrastructure bets by liquidating human capital. And some of it — perhaps the largest share — is Leadership Laundering: the settling of old organizational debts with a new excuse.

The urgency is this: the three faces require three entirely different responses, and the people running these organizations have every incentive to keep them confused.

If it is the Displacement Story, the response is retraining and transition support — new skills for a changed economy. If it is the Capex Trade, the response is accountability for what is being liquidated and what the actual return calculation looks like over ten years, not two. And if it is Leadership Laundering, the response is the one that nobody in a corner office wants to hear: you have been mismanaging human assets for decades, and you now have a technology story sophisticated enough to bury the evidence.

I have seen what happens when you go the other direction. At Royal Enfield, I inherited a workforce operating under low expectations, minimal engagement, and the quiet assumption that these were not high-capability people. We did not replace them. We deployed them. Profits grew twentyfold. The intelligence was there. It was always there. It was already paid for.

The tragedy of this moment is not that machines are becoming more capable. That is real, and it will continue, and managed well, it is not a disaster. The tragedy is that we are using the arrival of one kind of intelligence as cover for our failure to deploy another kind — the kind that has been showing up to work every morning for decades, waiting to be asked.

• • •

One number from Friday should stop every executive citing AI in their next layoff announcement.

Ninety percent of executives say AI has not affected hiring decisions. Ninety percent. The industry billing itself as the greatest disruption to human labor since the Industrial Revolution — and nine out of ten people doing the actual hiring and firing say it hasn’t changed their decisions in three years.

Either AI is radically less transformative than its promoters claim — which would make the displacement narrative a trillion-dollar exaggeration. Or the executives are lying about the reason — which would make it something considerably worse.

Both explanations are in the data. Both are probably partially true. Neither is comfortable. And neither will be featured prominently in the next earnings call that announces headcount reductions “in alignment with our AI-first strategy.”

Some of us are watching. Some of us have been watching for a long time. And we know the difference between a technology story and a leadership story.

This is a leadership story. It has always been a leadership story.

The workers losing jobs in February did not become redundant because machines got smarter. They became vulnerable because the organizations they gave their years to never got serious about deploying what those people actually knew.

That is not a disruption. That is a choice. And the people making it are hoping you will blame the algorithm.

Dr. Venki Padmanabhan is Plant Manager at Advanced Drainage Systems with 36 years of manufacturing leadership experience across three continents. He previously served as COO/CEO at Royal Enfield and COO at Ather Energy. His book, Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away, is forthcoming. Subscribe to The Long Game at thelonggameforall.substack.com.

Nick Lichtenberg did me an extraordinary kindness in Fortune this week. In a piece called “AI Just Gave You Six Extra Hours Back. Your Boss Already Took Them,” he quoted a plant manager from Ohio alongside Google directors, McKinsey partners, the chair of KPMG, and Baroness Dambisa Moyo of the House of Lords. I want to honor that generosity by doing what any good frontline worker does when someone hands them an opening — walk through it and get to work.

Nick assembled a magnificent chorus of voices all circling the same question: artificial intelligence is compressing tasks that once took days into minutes, so what happens to the hours that are left? A two-week audit at AES is now done in sixty minutes. Fifty percent of Google’s code is written by AI. A KPMG partner’s meeting prep dropped by seventy-five percent. The Dun & Bradstreet CTO, Mike Manos, put it with admirable candor: “I got the eight hours to two hours, but now I can get twenty hours of work.”

Every answer in the piece is about speed. I want to offer an answer about sight.

The Word That Caught Me

Manos used the phrase “widening aperture” to describe what AI makes possible — more problems to solve, more projects to chase, a bigger version of the job. It’s a photographer’s word. An aperture is the opening in a lens that determines how much light gets in, and therefore how much of the world the camera can see.

Manos means throughput. More light, more volume, more work crammed through the pipe.

But there is another way to read that word. What if the aperture that needs widening isn’t the workload? What if it’s the worker’s field of vision?

This is what I have spent thirty-six years studying, from General Motors to Royal Enfield to the corrugated pipe factory I run today in Wooster, Ohio. The single most consequential decision any company makes is not which technology to deploy. It is how much of the business it allows the frontline employee to see.

Three Openings

When I talk about widening the aperture for the frontline employee, I mean three specific expansions of sight. Not metaphors. Architecture.

First, see the whole business. Not just your station, your cell, your eight-foot stretch of the line — but the system. How the product reaches the customer. How the order triggers the production run. How quality at your hands becomes reputation in the market. In most American manufacturing plants, this view is systematically withheld. The worker is given a task, a rate, a shift. The aperture is kept narrow by design.

Second, see how money is made. Most plant workers have never seen a P&L. They have never been told what margin means, what their line contributes to revenue, what scrap actually costs in dollars the company will never recover. Open this aperture and suddenly the worker isn’t a pair of hands. They are an economic actor who understands the consequences of their own judgment.

Third, see your own reward. Not a generic bonus. Not profit-sharing that feels like weather — good years, bad years, who knows why. A fixed base that says we value your presence — and a variable component, above that base, that says we will pay you what you’re worth. Worth measured not in hours logged or widgets counted, but in Capability Capital: the deployed intelligence, judgment, and problem-solving that only a human being with a wide aperture can deliver. The variable line runs directly from “I saw this problem, I applied my intelligence, I solved it” to “here is what that was worth to the enterprise.” That is getting paid what you’re worth — and worth is measured in the capability you bring, not the time you spend.

When all three apertures are open — when the worker sees the business, understands the economics, and can trace their variable reward back to their own deployed Capability Capital — something happens that no amount of AI acceleration can produce on its own. You get a human being who is invested. Not managed. Not monitored. Invested.

That is where the ministry manifests.

The Suppression

So why don’t more companies do this? The standard excuse is that frontline employees aren’t capable of understanding the business. They lack the education. They lack the context. They wouldn’t know what to do with the information.

I have heard this excuse on four continents, in seven languages, across three decades. And I will tell you what it actually is.

It is fear.

Not fear that the worker can’t comprehend the business. Fear that the worker will comprehend the business. Because a worker who sees the whole system — who sees the margin, who sees what their labor actually generates, who sees the delta between what they produce and what they are paid — is a worker who can see the extraction. And visible extraction has to be justified. And justified extraction eventually has to be shared.

The narrow aperture is not a training failure. It is a design choice. The frontline is kept in the dark not because they can’t handle the light, but because capital cannot handle being seen.

Think about how perfectly this explains the pattern in Nick’s Fortune piece. Every executive he quotes is talking about what to do with the time AI freed up. Not one of them is talking about showing the worker where the value went. Manos says he can now get twenty hours of work out of what used to take eight. Growing for whom? The KPMG chair says “my business should be growing, and will grow.” He doesn’t say who shares in that growth. He doesn’t have to. The aperture is narrow enough that no one on the frontline can do the arithmetic.

The Harvard Business Review study Nick cites found that AI early adopters experience work as more intense — a phenomenon Boston Consulting Group researchers have taken to calling “AI brain fry.” Of course they do. They’re running faster on a treadmill they cannot see the end of, producing value they cannot trace, for a reward structure that hasn’t changed. The aperture of the task got wider. The aperture of understanding stayed shut.

The Proof That’s Been Hiding in Cleveland

For anyone who thinks what I’m describing is utopian, I have two words: Lincoln Electric.

Lincoln Electric is a welding equipment manufacturer headquartered in Cleveland, Ohio — sixty miles from my plant in Wooster. It has been in continuous operation since 1895. It has not laid off a single employee for economic reasons since at least 1948, and possibly since 1925. It has paid an annual profit-sharing bonus to every employee, without exception, every single year since 1934. That is ninety-one consecutive years of shared prosperity.

And it is ferociously, ruthlessly competitive. Lincoln Electric drove General Electric out of the welding industry. Its production employees are among the highest-paid industrial workers in Cleveland. Its revenue exceeds $3.8 billion. This is not a soft company coasting on goodwill. This is a company that wins — and wins because of how it treats its people, not despite it.

How? The aperture is wide open. All three openings, built into the architecture from the beginning.

Lincoln’s frontline workers see the business. An employee advisory board has met with the president every two weeks since the company’s early days. Open-door policy. Complete transparency on how piecework rates are set — and once set, those rates are guaranteed forever. The worker has the right to challenge every change.

Lincoln’s frontline workers see how money is made. They understand that when there is no profit, there is no bonus — and that when profit grows, the bonus pool has no upper limit. The company sets aside roughly a third of pre-tax profits for the annual bonus. In recent years, that pool has exceeded $100 million. The average bonus has run around $33,000 per employee — on top of wages that already lead the market.

Lincoln’s frontline workers see their own reward. The piecework rate is the fixed base — it says your presence and your labor have value. But the bonus is the variable, and it is calibrated to Capability Capital. Every employee is individually evaluated twice a year on four dimensions: output, quality, dependability, and — this is the one that matters — idea generation and cooperation. Not just how fast your hands move. How well your mind works. The worker who sees more, thinks more, and solves more gets paid more. They are getting paid what they’re worth — and worth is measured in deployed intelligence, not hours on the clock. The line from capability to reward is direct, legible, and real.

Since 1955, the average annual bonus has run at roughly seventy-seven percent of base pay. Read that again. Lincoln Electric workers routinely take home nearly double what they’d earn at any comparable manufacturer — and the company is more profitable for it, not less.

The Recession Test

But does it hold under pressure? In 1982, Lincoln Electric’s revenue dropped forty percent — from $450 million to $220 million — as every one of its traditional markets collapsed simultaneously. At any other company, that’s a layoff of a third of the workforce, minimum. Every CEO in Nick’s Fortune piece would have started cutting heads.

Lincoln didn’t lay off a single person.

They reassigned about fifteen percent of production workers to maintenance, to clerical work, to sales. Average pay was roughly halved. Executive compensation was cut too. Everyone absorbed the shock together. And the company stayed profitable. And it still paid the bonus.

When the recovery came, Lincoln didn’t have to rebuild institutional knowledge from scratch. It didn’t have to recruit and retrain. The intelligence was still there. Every worker who had been reassigned came back to the line with a wider view of the business than they’d had before — because they’d seen other parts of it during the downturn. The aperture had actually expanded through the crisis.

The well didn’t run dry. It deepened.

Amberg’s Twenty-Year Witness

Lincoln Electric is the American proof. Let me offer a European one.

Siemens operates a factory in Amberg, Germany, that kept the same 1,100 employees for twenty years while technology evolved around them. Those workers were not replaced by automation. They were formed alongside it. And over those two decades, they generated eight times the business output.

Same people. Same headcount. Eight times the result. Not because the machines got faster — though they did — but because the humans got wider. Their aperture expanded year by year. They understood the system. They understood the economics. They understood their role in an enterprise that was growing because of them, not in spite of them.

Siemens calls Amberg its “factory of the future.” I call it the factory that took the aperture seriously.

The Order of Operations

Here is where I part company with nearly every voice in Nick’s article.

The entire conversation — at Google, at McKinsey, at KPMG, at Dun & Bradstreet — starts with the tool. We deployed AI. We got these efficiency gains. Now what do we do with the humans?

That is the wrong sequence. Intelligence deployment precedes AI application. You widen the aperture first. You invest in the frontline worker’s ability to see the business, understand the economics, and connect their contribution to their reward. Then you hand them AI — not as a surveillance instrument or a speed multiplier, but as an amplifier of judgment they already possess.

A worker with a narrow aperture plus AI equals faster extraction until the well runs dry. The company milks the economic value of the knowledge that AI captured from past practice for maybe ten or fifteen years. But there is no new knowledge being developed, because humans develop knowledge. And then the well runs dry.

A worker with a wide aperture plus AI equals compounding intelligence. Because that worker knows what to ask the AI. They know what to question in the AI’s output. They know what the AI is missing — the tacit knowledge of the line, the customer, the slight change in the sound of the extruder that means the temperature is off. AI has no nose. It has no fingertips. It has no thirty years of pattern recognition earned in the smell of resin and the feel of a pipe wall. But it has tremendous analytical power — and in the hands of a worker who can see the whole business, that power compounds rather than extracts.

Lincoln Electric understood this before the word “artificial intelligence” existed. James F. Lincoln’s foundational insight, laid out in his 1951 book Incentive Management, was that workers suppress their own productivity when they fear it will cost them their jobs. The narrow aperture isn’t just intellectually confining. It is economically rational for the worker. If you can’t see where the value goes, and you suspect it isn’t coming back to you, why would you give your best thinking to the enterprise? You’d be funding your own replacement.

Lincoln’s answer was the Guaranteed Continuous Employment Plan. Your job is secure. Now give me your real intelligence. And the workers did. And the company became the most dominant force in its global industry for the better part of a century.

What Keynes Couldn’t See from Cambridge

Baroness Dambisa Moyo raised a haunting reference in Nick’s piece. John Maynard Keynes predicted in the 1930s that by 2030, technology would make a fifteen-hour workweek possible. Then he asked, with obvious anxiety, what people would do with all that free time. “Will they be contemplating God?” Moyo noted, adding her own worry about rootless young men around the world who are “not contemplating God in the manner in which we would want them to.”

I understand the anxiety. But I think Keynes — and Moyo — are asking the wrong question, because they’re imagining the freed hours as empty hours. Leisure. Void. Time without purpose.

The Ministry of Manufacturing offers a different answer. The freed hours are not free time. They are formation time. Time to learn the business. Time to understand the economics. Time to develop the judgment that makes a frontline worker not just fast but wise. Sanctuary — the safety to learn without fear. Ascension — the structured pathway to greater capability. Crucible — the real challenges that turn knowledge into mastery.

The workers at Siemens Amberg are not contemplating God in their liberated hours. They are doing something Keynes never imagined from the remove of Cambridge — they are growing into the technology rather than being replaced by it. And the workers at Lincoln Electric have been doing it since 1934, in Cleveland, in manufacturing, in the very heart of the Rust Belt that everyone else has written off.

Watch What Happens

So here is my answer to the question Nick’s article poses. AI freed up six hours. The boss took them. What now?

Open the aperture. Honestly. All three openings. Let the frontline employee see the business — the whole business, not the sanitized version. Let them see how money is made — the real numbers, the real margins, the real cost of the problems they solve every day. Let them see their reward — variable, specific, tied directly to their contribution, above and beyond base compensation.

Then hand them AI. Not as a replacement. Not as a treadmill set to a higher speed. As an instrument worthy of the musician.

And then watch what happens.

Watch the discretionary effort that no monitoring system can compel. Watch the problem-solving that no algorithm can replicate. Watch the compounding intelligence of a workforce that is invested — not because they were given a motivational speech, but because they can see the return on their own deployed capability.

Lincoln Electric has been watching it happen for ninety-one years. Siemens Amberg watched it happen over twenty. The companies that understand how to unlock this intelligence, engage their people, deploy the tacit knowledge they already have, and then layer AI on top? They are going to win extraordinarily.

The companies that simply cut — that keep the aperture narrow, harvest the AI efficiency gains, and send the savings to the balance sheet — will milk the well for ten or fifteen years. And then it will run dry. Because the humans who develop new knowledge will be gone. And AI, for all its magnificent speed, cannot develop what it has never seen.

The question isn’t whether AI gives you back six hours. It’s whether you have the courage to let your people see what those six hours are worth.

Venki Padmanabhan is a plant manager, former CEO, and the author of the forthcoming book Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away. He writes The Long Game on Substack and is co-founder of the Capability Capital Institute.

Venki Padmanabhan • The Long Game

This morning I called my mother in Vellore on WhatsApp. Forty-five minutes, Wooster, Ohio to Tamil Nadu, India. Her voice was clear. She told me about the temple visit, asked about the children, complained about the heat. A normal Saturday morning call between a son and his mother, separated by eight thousand miles and fifty years of migration.

It cost me nothing.

This afternoon, Reuters reported that Meta—the company that carried her voice to me—is preparing to lay off twenty percent of its workforce. Roughly sixteen thousand people. The same company that made my free call possible is now eliminating one-fifth of its employees to fund a six-hundred-billion-dollar bet on artificial intelligence. Mark Zuckerberg, Meta’s chief executive, has explained the logic plainly: “Projects that used to require big teams can now be accomplished by a single, very talented person.”

I sat with the phone in my hand afterward and did the math. In 1985, when I arrived in Pittsburgh as a twenty-five-year-old doctoral student in the Department of Industrial Engineering at the University of Pittsburgh, with a suitcase and an ambition, the same call would have meant dialing 011-91 on a rotary phone and running through AT&T’s international trunk lines at roughly $2.50 a minute. Forty-five minutes: $112.50. My teaching assistantship paid for tuition and $700 a month. One call to my mother would have cost sixteen percent of my monthly income. I would have written a letter instead. Or called for five agonizing minutes on a Sunday night, when the rates dipped, and listened to her ask if I was washing my clothes regularly and whether I had called my uncle in Flushing—$2.50 a minute for a mother’s inventory of concerns—and spent the rest of the week composing in my head all the things I didn’t have time to say.

So this morning’s call was a gift. A miracle of technology. A triumph of progress.

And yet something sat wrong in my stomach. A pit feeling. Because in thirty-six years of manufacturing, I have learned one unshakeable law: nothing is free. Every transaction has a cost. If you cannot see the cost, it means someone has designed the system so that you won’t.

What I realized, sitting with Zuckerberg’s layoff announcement in one hand and the residual warmth of my mother’s voice in the other, is that the same logic connects all of it—the free call, the invisible invoice, and the sixteen thousand people about to lose their jobs. It is a single system of extraction, operating in stages. And I have been watching it my entire career.

Stage One: Extract from the Customer

Meta reported $164 billion in revenue last year. WhatsApp has over two billion users. The service is free. The arithmetic should trouble anyone who has ever read a balance sheet: two billion people producing zero revenue each do not generate $164 billion. The gap between zero and $164 billion is you. It is me. It is my mother in Vellore, who has never once been asked what her data is worth or whether she consents to its extraction.

Here is what my forty-five-minute call actually provided Meta, at no charge and with no negotiation: confirmation of an international family network spanning the United States and South India. The frequency, duration, and timing of our contact patterns. The emotional cadence of a diasporic relationship—regular, sustained, likely involving an elderly dependent. My location. Her location. The full topology of my contact list, including people who have never installed WhatsApp and never agreed to anything. All of this was captured, correlated with my activity across Instagram and Facebook, and fed into an advertising profile that allows Meta to sell access to me—and to people like me—at $15 to $30 per thousand impressions.

Meta may have a bone to pick with my family. My wife forbids me from preening on Facebook. My daughter forbids my joining Instagram. WhatsApp is the last channel they have into me—and still, from that single thread, they can pull more than enough to build a profile, price it, and sell it. Imagine what they harvest from someone who gives them all three.

The 1985 AT&T call was expensive. But it was honest. I paid $2.50 a minute and AT&T provided a service. The transaction was legible. Both parties knew the price, the product, and the terms. The 2026 WhatsApp call was free. But it was opaque. I paid with behavioral data whose value I will never see, consented to terms buried in a seventy-five-page legal document designed to satisfy regulators rather than inform users, and contributed to a revenue stream from which I will never receive a single cent. The product changed from the call to the caller.

This is the first stage of extraction: harvest the customer.

Stage Two: Extract from the Worker

The moment I named what Meta was doing, I recognized it. I had been watching the same extraction my entire career. The venue was different. The mechanism was identical.

I have spent thirty-six years on factory floors—General Motors, Chrysler, Mercedes-Benz, Royal Enfield, and now Advanced Drainage Systems. On a production line, a worker’s hands perform the assigned task. But her mind does something far more valuable: it recognizes patterns. She notices that a particular torque sequence produces fewer rejects. She sees that a material batch from one supplier behaves differently than the same specification from another. She develops an intuition for when a machine is drifting toward failure—an intuition built from thousands of hours of sensory data that no sensor array can replicate. This is intelligence. It is deployed every shift, on every line, in every plant in the world.

And it is extracted for free.

When that worker’s observation leads to a process improvement, the improvement is captured as a management initiative, an engineering change order, a “kaizen event” attributed to the system rather than the person. The margin improvement flows to the income statement. The stock price reflects it. The worker receives her hourly wage—the same wage she would have earned if she had noticed nothing, contributed nothing, and kept her intelligence to herself.

The architecture is identical to Meta’s. Make the contribution invisible to the contributor. Aggregate it at scale. Monetize it through channels the contributor never sees. Whether the intelligence is a frontline worker’s quality instinct or a WhatsApp user’s behavioral data, the extraction mechanism is the same: value is generated by one party, captured by another, and the person who generated it has no seat at the table where the price is set.

This is the second stage: harvest the worker.

The Bauble Economy: Extraction Through Employment Itself

But the extraction at Meta goes beyond harvesting existing workers. There is a stage between harvesting and discarding that no one is naming. Call it the Bauble Economy: the cycle of speculative hiring that treats human beings as venture bets.

The numbers tell the story with brutal clarity. In March 2020, Meta employed 48,268 people. By September 2022, that number had swelled to more than 87,000—nearly doubling in two and a half years. The reason was the metaverse. Zuckerberg, freshly rebranding the company from Facebook to Meta, announced plans to hire 10,000 employees in Europe alone to build his vision of an immersive internet. The company poured resources into Reality Labs, the division tasked with making the metaverse real. People were recruited from across the industry, offered lavish packages, relocated across countries. They rearranged their lives for a vision.

The vision lost ninety billion dollars.

Reality Labs has accumulated approximately $90 billion in cumulative operating losses since late 2020. The metaverse generated roughly one percent of Meta’s total revenue despite consuming billions annually. In November 2022, Zuckerberg laid off 11,000 people—thirteen percent of the company. He declared 2023 the “year of efficiency.” Four months later, he cut another 10,000. Headcount fell twenty-two percent in a single year, from 86,000 to 67,000. Then the company began hiring again—this time for AI. By the end of 2025, headcount had climbed back to 79,000. Now comes the next twenty percent cut.

Read the sequence again: 48,000 to 87,000 for the metaverse. Down to 67,000 when the metaverse failed. Back up to 79,000 for AI. Now down again to perhaps 63,000 as AI requires data centers instead of people. Each cycle, tens of thousands of human beings uproot their families, build expertise in a domain the company will abandon within thirty-six months, and are then told their labor is no longer aligned with the strategic direction. The unvested equity they were promised evaporates. The mortgage they took out based on a compensation package that assumed continued employment comes due on a single income. The institutional knowledge they developed—about what worked, what didn’t, why the metaverse couldn’t find its users—walks out the door with them, unasked for and unrecorded.

Alphabet runs the same play. Google added nearly 72,000 employees over three years from 2020 through early 2023, increasing headcount by thirty-eight percent. Then Sundar Pichai announced 12,000 layoffs, explaining that they had hired for “a different economic reality.” Between 2023 and 2025, Google cut an estimated 15,000 to 20,000 positions. The strategic rationale shifted from post-pandemic correction to AI optimization. Different bauble. Same discarded workforce.

A manufacturer who operated this way would be bankrupt within two cycles. You cannot retool a production line every eighteen months, fire the workers who understood the old line, hire new ones who don’t yet understand the new one, and expect quality output. The learning curve alone would destroy you. But technology companies can absorb this chaos because their revenue comes from legacy products—Facebook, Instagram, WhatsApp, Google Search, YouTube—that print money regardless of how badly management misallocates the current workforce. The advertising cash cow subsidizes the irresponsibility. The employees absorb the cost.

And here is the point that connects the Bauble Economy to the broader extraction system: none of these people were formed. They were acquired as capabilities, pointed at a project, and when the project failed, they were written off like depreciated equipment. No one asked what they learned. No one captured the institutional knowledge about why the metaverse couldn’t find its market. No one asked whether the models failed partly because the intelligence of 40,000 people was never actually activated—just directed. The extraction model does not just rob frontline workers of their intelligence. It robs knowledge workers of their formation. It treats the most educated workforce in history the same way it treats the production floor: as raw material to be used and discarded when the next shiny bauble catches the CEO’s eye.

This is the half-stage between harvesting and discarding: extraction through employment itself.

Stage Three: Discard Both

This is what Zuckerberg’s latest layoff announcement represents. Once you have extracted all the intelligence you can from human contributors and encoded it into models, the logical terminus of the extraction economy is to remove the human from the loop entirely.

“Projects that used to require big teams can now be accomplished by a single, very talented person.” That is not a prediction. It is a confession. The extraction is complete when the source is no longer needed.

Consider the full picture at Meta. The company reported 79,000 employees at the end of 2025. It is preparing to cut twenty percent of them while simultaneously spending $600 billion on data centers and paying elite AI researchers employment packages that can add up to hundreds of millions of dollars. Capital floods upward toward infrastructure and a tiny talent aristocracy. The broad workforce gets zeroed out.

And this comes after a year of technical stumbles. Meta’s Llama 4 models underperformed. Plans for its larger model, Behemoth, were shelved. A newer model called Avocado reportedly failed to meet expectations. The response to these failures was not to invest in the humans who might diagnose what went wrong. The response was to double down on infrastructure spending and cut more people. No one in Menlo Park appears to be asking whether the models failed because of how Meta treats human intelligence in its pipeline.

The logic is perfectly consistent and perfectly destructive: harvest the customer’s data for free. Harvest the worker’s intelligence for free. Churn through employees on speculative bets, discarding them when the bet fails. And when the harvesting is complete, remove the human from the operation entirely.

Harvest the customer. Harvest the worker. Discard both.

Others Have Seen the Pieces

Shoshana Zuboff named the digital half of this extraction with devastating precision: a system that claims human experience as free raw material for hidden commercial practices of extraction, prediction, and sales. Jaron Lanier approached it from the economics of dignity, arguing that digital information is really just people in disguise. Antonio Casilli and the digital labor school drew the line between factory and platform most directly.

These thinkers are right. But they all approached the problem from the digital side, looking backward toward industrial analogies. None of them started where I start—on the factory floor, watching a second-shift operator solve a problem that will save the company $40,000, and then clock out at the same wage as the operator next to her who solved nothing. And none of them stayed long enough to see the full arc: the harvesting, the churning, and the elimination of the contributor altogether.

The extraction is not an analogy. It is the same system, operating across every domain.

I Have Seen the Alternative

I know this system can work differently because I have been inside the alternative.

In the late 1990s, Royal Enfield was nearly dead. The Indian motorcycle company, then a subsidiary of Eicher Motors, was producing about two thousand bikes a month from a single aging factory in Tiruvottiyur, Chennai. The bikes leaked oil. Quality was abysmal. Losses were mounting. The chairman wanted to shut the brand down entirely. His son, Siddhartha Lal, asked for two years to turn it around.

What followed was not an exercise in extraction. It was an exercise in cultivation.

The workforce was not dismissed as incapable. It was redesigned around. The system was rebuilt to recognize, amplify, and reward the intelligence that workers were already deploying every shift. Quality systems were redesigned not to catch defects but to prevent them—by trusting frontline knowledge. The factory floor was treated not as a cost center to be minimized but as an intelligence network to be activated.

The results were not incremental. By the time I had served my turn there, the same workforce that had been written off had helped produce a twenty-fold increase in profitability. And the company kept growing. Royal Enfield opened a second manufacturing facility in Oragadam in 2013 and a third in Vallam Vadagal in 2017. Sales rose from those two thousand bikes a month to over a million motorcycles in 2025—a forty-fold increase in volume. The workforce grew from a skeleton crew in a single leaking factory to roughly fifteen to nineteen thousand employees across three modern plants and operations in more than sixty countries.

Let me say that again, because it is the direct negation of the Zuckerberg thesis: Royal Enfield did not grow by eliminating its workers. It grew by amplifying them. It did not replace big teams with single talented individuals. It made its teams more talented. Capital and labor ascended together. I call it the Twin Helix.

The Minnow and the Behemoth

I can hear the objection already. Royal Enfield is a minnow. Fifteen thousand employees and a million motorcycles is a rounding error next to Meta’s $1.8 trillion market cap. The same will be said of the other companies where I have seen the cultivation model work: Siemens’s Amberg factory, which produces fourteen million automation components a year at a quality rate of 99.9989 percent with a workforce that has been continuously upskilled over decades. Lincoln Electric, which has operated a no-layoff policy since the 1940s and consistently outperforms its competitors in a commodity industry. These are minnows, the tech establishment will say. Physical products require physical hands. Software is weightless. AI changes everything. You are comparing apples to assembly lines.

But the objection answers itself. The fact that the cultivation model has only survived at companies small enough, or privately held enough, or culturally rooted enough to resist Wall Street’s quarterly extraction logic—that is not a failure of the model. It is an indictment of the system. The public markets reward the hire-fire bauble cycle and punish patience. Meta’s stock price jumped when it announced metaverse budget cuts. Every layoff announcement is greeted by analysts as discipline. Every investment in workforce formation is questioned as overhead. The minnows are not small because the model fails at scale. They are small because the financial system selects against them.

And let us examine the behemoths honestly. Are they actually performing? Meta has burned $90 billion on the metaverse, fired and rehired tens of thousands of people in overlapping cycles, and still generates ninety-nine percent of its revenue from products built over a decade ago—Facebook, Instagram, WhatsApp. Google added 72,000 people, cut 20,000, and its core revenue still comes from Search, a product essentially unchanged since 2004. These companies are not innovating with their workforce churn. They are churning in place while legacy products print money. The behemoths are not laughing from a position of operational excellence. They are laughing from a position of monopoly rent.

Royal Enfield grew its workforce alongside its output for twenty-five years and just crossed a million motorcycles. Meta doubled its workforce in two years, burned ninety billion dollars, and is now on its third cycle of mass layoffs. Tell me again which model does not scale.

The Incomplete Scoreboard

They will say the extraction model works. And by the only metric anyone currently reports, they are right. Meta’s market cap is $1.8 trillion. Alphabet’s is $2 trillion. Zuckerberg’s personal net worth exceeds $200 billion. By the scoreboard we have, they are winning.

But the scoreboard is incomplete. It measures wealth concentration, not wealth creation. It measures what accrued to the top, not what was generated across the whole system.

What if we placed next to every market cap figure the median net worth trajectory of the people who built that value? Not the founders. Not the board. The engineers and designers and content moderators and operations staff. The 40,000 people hired for the metaverse and discarded when the vision changed. The 16,000 about to be shown the door to make room for data centers. What happened to their household wealth? Their mortgage qualifications? Their unvested equity that evaporated when they were shown the door? Their children’s college funds tied to stock options priced at the hiring peak?

Nobody tracks that number. Nobody reports it. No CNBC ticker runs it. No analyst asks about it on the earnings call.

Now imagine the same measurement at Royal Enfield. Fifteen to nineteen thousand employees, many of whom have been there for a decade or more—compounding skill, compounding seniority, compounding stability. A workforce that grew alongside the business for a quarter century has a very different wealth trajectory than a workforce that was hired in 2021, fired in 2023, and is now watching from the outside as the stock they were promised vests without them.

My cousin Dr. Sridhar Ramamoorthi likes to invoke a line often attributed to Einstein but actually penned by the sociologist William Bruce Cameron: “Not everything that can be counted counts, and not everything that counts can be counted.” That single sentence explains why extraction persists and cultivation doesn’t. The value a worker carries in her head cannot be counted. So it doesn’t count. The prosperity a business model creates—or destroys—across its entire workforce cannot be counted. So it doesn’t count. Until she’s gone, and suddenly no one can figure out why the line won’t run right. Until they’re all gone, and no one can figure out why the models keep underperforming.

This is precisely why Dr. Ramamoorthi and I co-founded the Capability Capital Institute: to build the metrics that make the invisible visible. Not just the value of deployed human intelligence on a balance sheet—though that is where we start—but the full prosperity picture. What does a business model do to the net worth, the stability, the formation, and the dignity of the people inside it? If we can make that as legible as market cap, the entire calculus changes. What Cameron lamented as uncountable, we intend to count.

The Scoreboard That Already Exists

I do not need to imagine what this scoreboard would look like. I work at a company that built one.

Advanced Drainage Systems, where I am a plant manager and where I write in a personal capacity, established an employee stock ownership plan in 1993. Everything that follows is drawn from the company’s SEC filings and published interviews—the same information available to any investor. Over three decades, the ESOP allocated approximately twelve million shares to workers—not executives, not founders, workers. People who make stormwater pipes. People who drive forklifts. People who run extrusion lines. When the ESOP converted in 2022, the stock was trading above $120. Today it trades near $170. The aggregate value of those shares in employee hands approaches two billion dollars.

I know people on the factory floor in Wooster, Ohio, right now who watch the stock price every morning. Not because they are day traders. Because every tick toward $200 changes when they can retire, whether they can help their grandchild with college, how much dignity the last chapter of their working life will hold. That is what shared prosperity looks like when someone bothers to build the mechanism.

Joe Chlapaty, the CEO who established this plan, was not a radical. He was a plastics manufacturer from Dubuque, Iowa, who believed that if you asked people to build a company, they should own a piece of what they built. His explicit goal was to provide jobs that enable employees and their families, as he put it, “to live as best we can a middle-class life.” He kept eighteen percent of the shares himself when he retired. He did not cash out and leave. He kept his wealth alongside his workers’ wealth.

Under Chlapaty’s leadership and that of his successor Scott Barbour, ADS grew from $1.26 billion in revenue and 4,500 employees to $2.9 billion in revenue and 6,000 employees. Margins doubled. The market cap went from $850 million at the IPO to over $13 billion. And the workers went with it—not as raw material being extracted, not as overhead being minimized, but as shareholders whose prosperity was structurally bound to the company’s. That is the Twin Helix made concrete. That is the alternative to the Bauble Economy. And it is not happening in a business school case study. It is happening in a stormwater pipe factory in Ohio while Mark Zuckerberg prepares his next round of severance letters.

Why This Isn’t the Norm

If the logic is this clear—and it is—then why doesn’t it govern every factory, every platform, every transaction?

Because financial markets punish patience. The cost of investing in frontline capability shows up this quarter. The return shows up in eighteen months. Wall Street sees the cost and does not wait for the return. The same temporal distortion governs platform capitalism: Meta’s stock price reflects the data it extracts today, not the trust it erodes over decades.

Because the managerial class—and the platform architect class—have been educated in a theology of labor as cost. Since the 1970s, the dominant mental model has treated human contribution as a depreciating line item to be minimized. Workers are overhead. Users are raw material. In neither case are they partners in value creation. This is not economics. It is ideology dressed as economics.

And because the people who benefit most from extraction have the most power to define what “normal” looks like.

The Call and the Calling

My mother is eighty-three years old. She does not know what metadata is. She does not know that her Sunday morning conversation with her son is a data point in a behavioral graph that spans two billion nodes. She does not know that the company carrying her voice is about to tell sixteen thousand families that their labor is no longer required. She knows that her boy called, that his voice was clear, and that the grandchildren are well. That is enough for her.

It is not enough for me.

Not because I begrudge Meta its infrastructure or its ingenuity. The technology that lets me hear my mother’s voice in real time across eight thousand miles is genuinely miraculous. But miracles should not require surrendering something you were never told you were giving away. And they should not require discarding the people who made the miracle possible in the first place.

In 1985, AT&T said: dial 011-91, pay us $2.50 a minute, and we will connect you to your mother. It was expensive and honest.

In 2026, Meta says: pay us nothing and we will connect you to your mother—and also harvest your behavioral intelligence, map your relationships, profile your identity, and sell predictions about your future actions to the highest bidder, in markets you will never see, at prices you will never know. And when we have extracted enough, we will begin removing the humans from our own operation, too.

The same exchange happens every morning in every factory in America. The worker clocks in, deploys intelligence, and clocks out at a wage that reflects her time but not her mind. The value she created is captured, aggregated, monetized—and she is told she should be grateful for the steady paycheck, just as I am told I should be grateful for the free call.

I am grateful. And I am unsatisfied. Because gratitude for a gift should not require blindness to an extraction.

What I want—for the worker on the line, for the user on the platform, for the sixteen thousand about to receive their severance letters, for the forty thousand who already received theirs when the metaverse dream dissolved, and for my mother on the other end of that WhatsApp call—is simple. Transactions where the benefit is commensurate to the intelligence deployed. Technology that amplifies human capability rather than automating it away. A financial system that measures prosperity where it is created, not only where it accumulates. And a world where capital and labor recognize that their fates are bound together in a helix that rises only when both strands are honored.

Zuckerberg believes the future belongs to single, very talented people aided by AI. I believe the future belongs to many talented people, aided by each other. I have seen both models. One of them built a million motorcycles by growing its workforce. Another turned pipe makers into shareholders and grew a $13 billion company without discarding the people who built it. The third has spent ninety billion dollars chasing baubles, discarded tens of thousands of lives in the process, and is preparing to do it again—all while living off products that were built a decade ago by people who are no longer there.

This is not a radical vision. It is the most logical norm there is. The fact that it must be argued for tells you everything about who currently writes the rules.

* * *

Venki Padmanabhan is a plant manager at Advanced Drainage Systems, a writer, and a founder of the Capability Capital Institute. He is the author of the forthcoming book Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away. He writes at thelonggameforall.substack.com.

You can watch the author express this essay on Youtube.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.

I know what some of you are going to say.

When I argue that the dominant manufacturing operating model systematically suppresses frontline intelligence — that there are millions of workers in American factories and hospitals, and warehouses and fulfillment centers whose problem-solving capability has been deliberately designed out of their jobs — some of you will call it romantic nonsense. You’ll say these are simple jobs performed by people the labor market sorted there. That there’s no hidden intelligence being suppressed. That the system is just efficient.

I welcome that objection. Because the evidence against it is overwhelming. And I’m going to spend the next eight weeks laying it out — not with theory, not with philosophy, but with data, natural experiments, and documented performance outcomes that the skeptics will have to reckon with.

We start with the strongest piece of evidence I know. A story so clean in its experimental design that if you proposed it as a controlled study, your IRB would reject it as too good to be true.

We start with NUMMI.

The Worst Plant in America

By the late 1970s, General Motors’ Fremont Assembly Plant in California had earned a reputation that even GM’s own leadership didn’t dispute: it was the worst plant in the system.

Absenteeism ran above 20 percent. On some mornings, so few workers showed up that supervisors had to cross the street to the local bar to drag people back to the line. Wildcat strikes were routine. Workers drank on the job, did drugs on the job, and — as multiple firsthand accounts confirm — had sex in the plant during shifts. Sabotage was a sport. Workers left bolts deliberately untightened. They stuffed Coke bottles inside door panels so the cars would rattle and annoy customers. Cars regularly rolled off the end of the line inoperable and had to be towed away for repairs.

Under the union contract, workers were virtually impossible to fire. Management and labor existed in a state of cold war. The plant was, by any reasonable measure, a disaster.

In 1982, GM shut it down.

If you stopped the story here, you’d have a clean narrative. Bad workers. Bad culture. Bad outcomes. The labor market sorted these people into a factory, they proved themselves incapable, and the rational response was to close the doors.

That narrative is wrong. And what happened next proved it.

The Experiment Nobody Would Have Designed

In 1984, Toyota and GM reopened the Fremont plant as a joint venture called New United Motor Manufacturing, Inc. — NUMMI. Toyota would manage the plant. GM would contribute the facility. Both companies would sell the vehicles produced there.

Here is what Toyota did that made this a natural experiment of extraordinary power:

They rehired the same workforce.

Over 85 percent of NUMMI’s initial employees were UAW members from the old GM Fremont plant. Not the good ones. Not a carefully screened subset. The same workforce — including, at the UAW’s insistence, the same union leadership that had presided over GM’s worst operation.

GM was against rehiring the militants. Toyota agreed to it anyway. They believed their system could work with any workforce — a hypothesis they were about to test with the hardest possible case.

Toyota did one other critical thing before reopening. They sent groups of these workers to Toyota’s Takaoka plant in Japan. Not for a PowerPoint presentation. Not for a classroom seminar. They put them on the assembly line. They let them work inside a system that treated every worker as a problem-solver, that gave them authority to stop the line when they saw a defect, that asked for their ideas and implemented them.

The workers came back to Fremont and started building cars.

The Results That Should Have Changed Everything

Within one year, the plant that had been GM’s worst became GM’s best.

Not good. Not improved. The best.

John Shook was there. An industrial anthropologist hired by Toyota in 1983 to help develop training programs for the venture, Shook had a front-row seat to what he later called one of the most striking stories of cultural reinvention he had ever witnessed. Writing in the MIT Sloan Management Review, he put the outcome in terms that leave no room for ambiguity:

“We took the quality of the plant from GM’s very worst to GM’s very best — not just bad to good, from worst to best — in only one year. The exact same workers, including the old troublemakers. The only thing that changed was the system. The production and management system.”

Let that sink in.

The workers who had been drinking on the line, sabotaging cars, walking off the job — those workers, in the same building, with the same union, produced the highest quality vehicles in General Motors’ entire North American operation. In twelve months.

A 2024 Harvard Business School case study by Willy Shih confirmed what practitioners had known for decades: Toyota transformed Fremont into the most productive auto assembly plant in the United States, with quality comparable to its Japanese factories.

Isao Yoshino, the Toyota training manager at NUMMI’s launch, reflected on the transformation with a simplicity that cuts through every layer of management theory: “We did not do anything else. We just offered the opportunity for them to learn, and they voluntarily decided to change themselves.”

What Actually Changed

If it wasn’t the people, what was it?

The operating model. Specifically, Toyota replaced a system designed to suppress frontline intelligence with one designed to deploy it.

Under GM’s old model, the Fremont plant operated on pure Taylorist principles — principles I’ll trace back to their explicit origins in the next essay. Workers performed narrow, prescribed tasks. They had no authority to stop the line. Their ideas were not solicited. Their judgment was not valued. Problems were management’s responsibility to identify and fix. The workers’ job was to execute.

Under this model, the rational worker response was exactly what GM got: disengagement, cynicism, sabotage, and substance abuse. Not because the workers were incapable, but because the system told them every day that their capability was unwanted.

Toyota’s model inverted every one of these assumptions. Workers were organized into teams with real problem-solving authority. Any worker could stop the production line — the famous andon cord — when they spotted a quality issue. Suggestions were actively solicited and rapidly implemented. Workers rotated through jobs and were cross-trained. The uniform, cafeteria, and parking lot were shared across all levels of the organization. A no-layoff policy removed the existential fear that had driven much of the adversarial behavior.

The cultural changes ran deeper than process. Workers who visited Japan identified what motivated them to change: they were treated as intelligent contributors whose judgment mattered. For many of them, it was the first time in their working lives that a manager had asked what they thought.

The Part Nobody Wants to Talk About

Here is what should haunt every manufacturing executive who reads the NUMMI story:

GM had a front-row seat for twenty-five years. They embedded managers at NUMMI. They studied the Toyota Production System up close. And for fifteen of those years, they failed to transfer the lessons to a single other plant in the United States.

It wasn’t that they didn’t try. GM sent a team to replicate the approach at their Van Nuys plant in Southern California. It failed. The workers there had never lost their jobs and didn’t believe they were at risk. But more importantly, themanagement system didn’t change. As one of the GM managers who attempted the transfer recalled, the problem was a resistance to change that went beyond the workers. It permeated the entire management structure.

By 1998 — fifteen years after NUMMI’s founding — GM had still not successfully implemented lean manufacturing across its U.S. operations. By 2009, the company was bankrupt.

The lesson is not that GM’s managers were stupid. The lesson is that intelligence suppression is structural. It’s embedded in the operating model, the management incentives, the labor relations framework, and the cultural assumptions about what workers are capable of. It’s not something one plant manager can fix by copying a set of tools. It’s a system — and as Deming told us, the system belongs to management.

What This Means for the AI Conversation

I opened this series by referencing a Forbes article about AI as a “Toolmaster” — a force multiplier for professionals who direct it with expertise and judgment. The article profiles marketing directors and lawyers and CFOs who use AI to amplify their already-deployed intelligence.

Beautiful. Also incomplete.

The article assumes — as nearly the entire AI conversation assumes — that the human intelligence is already active. That the professional already has judgment, context, and domain expertise. AI just makes them faster.

What about the 80 million Americans who work frontline jobs in manufacturing, healthcare, logistics, retail, and service? What about the workers whose problem-solving muscles have been atrophied by a century of Taylorist operating models that told them, explicitly, to stop thinking?

You cannot hand a power tool to someone whose intelligence has been systematically suppressed and expect Toolmaster results. The AI becomes another compliance instrument — or, worse, another justification for eliminating the human entirely.

NUMMI proved that the intelligence is there. It proved that the same people, in a different system, will generate extraordinary outcomes. The question for the AI era is not whether frontline workers are smart enough to use these tools. NUMMI answered that forty years ago.

The question is whether management is brave enough to build the system that lets them.

Next week: “The Confession in the Blueprint” — Frederick Taylor told us exactly what he was doing. We just stopped reading.

Dr. Venki Padmanabhan is a plant manager with 36 years of global manufacturing leadership experience, including executive roles at GM, Chrysler, Mercedes-Benz, Royal Enfield, and Ather Energy. He holds a PhD in Industrial Engineering from the University of Pittsburgh. He is the author of the forthcoming Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away and co-founder of the Capability Capital Institute. He writes The Long Game at thelonggameforall.substack.com.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.

Dr. Venki Padmanabhan | The Long Game | March 2026

For The Long Game for All (thelonggameforall.substack.com)

∗ ∗ ∗

On Thursday, Amazon’s retail website locked shoppers out of checkout for six hours. Not a DDoS attack. Not a server fire in Northern Virginia. An AI agent followed advice from an outdated internal wiki, and the system that serves roughly 310 million active customers went dark in the middle of the afternoon.

The same week, Jensen Huang stood on a stage and told the world not to ask for ROI on AI. “Let a thousand flowers bloom,” he said, borrowing Mao’s most famous invitation to chaos.

I want to propose a different metaphor. Not a thousand flowers. One bloom. Specific, singular, already planted, already paid for—and, in Amazon’s case, already ripped out of the ground to make room for the machine that just failed.

∗ ∗ ∗

What Actually Happened

According to the Financial Times, an internal document prepared for Amazon’s emergency review identified “GenAI-assisted changes” as a contributing factor in a pattern of incidents stretching back to Q3 2025. Four high-severity failures hit the retail site in a single week. The internal document was then edited—the AI reference deleted—before the meeting took place.

Amazon’s public response was revealing in what it conceded while trying to minimize. The company said only one incident involved AI tools, and that the cause was “an engineer following inaccurate advice that an agent inferred from an outdated internal wiki.”

Read that sentence again. An agent inferred advice from an outdated wiki.

Someone used to keep that wiki current. Someone used to know, from years of working with the system, which pages were stale and which were gospel. Someone used to be the institutional memory that prevented a new engineer from following a three-year-old procedure on a live production environment.

That someone was already paid for. And Amazon cut them.

∗ ∗ ∗

The Pattern Nobody Wants to Name

Amazon laid off roughly 14,000 corporate workers in October 2025—mostly middle managers, the people who carry institutional knowledge in their heads the way a root system carries water. Another 16,000 followed in January. That’s on top of 27,000 cut between 2022 and 2023. CEO Andy Jassy wrote in an internal memo that the company would need “fewer employees thanks to AI-driven efficiency gains.” When the October cuts came, he reframed the rationale as “culture”—the company had grown too fast, needed to be “lean.”

(There’s that word again. Lean. The word manufacturing has been hiding behind for thirty years when it means: we cut the people and kept the machines.)

But a separate Amazon memo announcing the same layoffs cited the need to adapt to “transformative technology.” The real reason. The reason they always walk back. The reason that now has a six-hour checkout outage attached to it.

Here is the pattern: fire the people who know things, replace them with systems that don’t, then act surprised when the systems infer from an outdated wiki because nobody is left to update it. It is not an AI failure. It is a bloom failure. Amazon ripped the living intelligence out of its own soil and expected artificial intelligence to photosynthesize in the dark.

∗ ∗ ∗

I Have Seen This Before

People will read this essay and wonder what a manufacturing guy knows about AI deployment at a trillion-dollar tech company. Fair question. Here is my answer: the pattern is identical. I have watched it play out not in server rooms but on assembly lines—and the economics of the failure are just as damning when you actually do the math.

During my Walden Pond years—two years on the third shift at GM’s Lansing Delta Township plant, launching three SUVs simultaneously—I witnessed the spare tire robot.

The context: every year, a plant manager mandates productivity improvements. The chain of command must go find a certain number of “heads” to remove from the line. Sometimes it’s a redistribution of work, taking waste out of movement. Sometimes it’s blunt: we’re bringing in a robot, here’s the number of people we no longer need. You can imagine that in a traditional GM-UAW setting with entrenched positions, the floor has zero interest in cooperating. And the shift leaders—guys like me—have every incentive to force it through, because our stock rises or falls on whether we deliver those numbers.

The business case looked clean. In the chassis area, downstream from my trim shop, a team member on each shift would take a spare tire, lift it with an assist device, and place it in the rear compartment. The proposal: install a robot to do the job. Eliminate one person per shift—roughly $120,000 all-in with benefits—times three shifts, $360,000 a year in savings. Total cost of the robot, programming, enclosure fabrication, MES integration, installation: roughly two million dollars. Payback in five to six years. Then you never have to hire those positions again. Sweet business case.

Here is what the business case did not include.

For the first three to four months, the chassis area suffered constant downtime. Programming bugs. MES integration failures. The robot would crash, the line would stop, and the entire plant—trim, chassis, four hundred people—would stand and wait. My trim shop would finish its work, fill every buffer, and then block against chassis with nothing to do. Hours and hours of it.

And here is the part nobody modeled: the floor offered no help. Not because they were lazy. Because their interests were perfectly misaligned. The robot was there to eliminate their colleagues. The skilled trades didn’t have knowledge of the new system. So the only people debugging it were the systems integrator and the robot installer—specialists who understood the machine but knew nothing about the application domain, going through failure mode after failure mode after failure mode with no deployed intelligence from the people who actually knew the process.

Then came the secondary failures—the ones no business case ever captures. In the old process, I would load a plastic trim panel at the end of my trim shop and let it ride through the chassis area until it was needed. Various team members along the way would reposition it depending on what they were doing in the rear compartment. At least once a night, a panel would end up in the wrong spot. The spare tire robot would come in, crush into it, crash, need to be homed, reprogrammed, restarted. Five to ten minutes every occurrence. This became a nightly blame game—chassis would say, “Trim, you shut us down.” We’d spend the rest of the shift arguing whether they were using us as a bogeyman to hide other problems.

Now do the real math. This was a sixty-jobs-per-hour plant. One car every minute. Every minute we didn’t build a car, we lost roughly $55,000 in revenue and perhaps $10,000 in profit. That means every hour of downtime cost $600,000 in lost profit—nearly two full years of the labor savings the robot was supposed to deliver. If somebody had done an honest, forensic accounting of the cumulative downtime during that robot’s installation—say, eighty hours over those first months, which is conservative—you are looking at $48 million in lost profit. Against a robot that cost two million and saved $360,000 a year. The downtime losses alone were twenty-four times the cost of the robot. That five-to-six-year payback the business case promised? It will never pay back. And I am not exaggerating for effect. I am describing what happens when you remove the bloom and then try to calculate the cost.

The question nobody asked was the one that mattered: Could the economic improvement we sought with the robot have been accomplished by deploying the intelligence of the people already on the floor—a line rebalance, a process kaizen, a genuine alignment of their interests with the company’s? The answer, almost certainly, was yes. But that path required trusting the bloom. And the institution had no framework for that.

This is exactly what happened at Amazon. Swap the spare tire robot for a GenAI agent. Swap the outdated wiki for a three-year-old procedure nobody maintained. Swap the skilled trades who wouldn’t help because their interests were misaligned for the middle managers who weren’t there because they’d been laid off. The topology is identical. The only difference is that Amazon’s failure happened to 310 million customers instead of four hundred workers on a night shift in Lansing.

∗ ∗ ∗

What I Mean by “The Bloom”

I have spent thirty-six years on factory floors across three continents—General Motors, Chrysler, Mercedes-Benz, Royal Enfield, Advanced Drainage Systems—and I have come to see a pattern so consistent it functions as a law: every organization sits on intelligence it has already bought and is actively suppressing.

The frontline worker who knows that Line 3 pulls left on humid days. The maintenance tech who can hear a bearing going bad two shifts before the vibration sensor catches it. The middle manager who remembers why we stopped using that vendor’s API in 2019. This intelligence is not mysterious. It is not hiding. It is standing right there on the floor, every shift, drawing a paycheck. Already hired. Already trained. Already paid for.

I call the system that protects and deploys this intelligence The Bloom. Not as metaphor—as mechanism.

A bloom is what happens when a bud is protected long enough to open. The lotus—my tradition’s signature flower—grows in mud, rises through murky water, and opens immaculate on the surface. The mud is not the obstacle. It is the nutrient. The factory floor is the mud. The worker’s tacit knowledge is the bud. And the structures that protect that bud until it opens are what I call the Three Opercula: Sanctuary (the company’s provision of what labor actually needs—Time, Love, Health, Wealth—so that workers can grow without existential fear), Ascension (the systematic pathway from task-holder to capability-builder, where people are measured on capability capital and paid based on what they learn and contribute as they rise), and the Sun(training and development in high-pressure contexts—the intense light that forces the bud to open into deployed intelligence). Remove any one and the bud rots before it blooms.

Beneath these protective structures lies what I call the Twin Helix—the intertwined strands of Capital’s goals (earnings, growth, innovation, brand) and Labor’s goals (time, love, health, wealth). Like DNA, the two strands are not opposing forces. They are complementary codes that, when read together, produce something neither creates alone. Every failed automation story I have ever witnessed—and I have witnessed many—comes down to the same error: someone pulled one strand out of the helix and expected the other to replicate on its own.

Amazon just demonstrated that error at scale, in public, to 310 million customers. The spare tire robot demonstrated it at smaller scale, in private, on a night shift in Michigan. The error is the same.

∗ ∗ ∗

Jensen Huang’s Thousand Flowers

On the same day Amazon was scrambling to explain why its AI broke the website, Jensen Huang was telling an audience that the path forward was to invest in AI without asking what it returns. Let a thousand flowers bloom, he said.

It’s a revealing choice of phrase. Mao’s original Hundred Flowers Campaign was an invitation for intellectuals to speak freely—followed by a purge of everyone who did. The thousand flowers rhetoric sounds like abundance. It is actually the language of expendability: plant so many that it doesn’t matter which ones die.

That is the opposite of what a bloom requires. A bloom is not scatter-and-pray. A bloom is specific. It is one bud, in one place, protected by structures that were built to help it open. It requires patience, sequence, and the recognition that the intelligence you need is already there—you just have to stop cutting it back.

Huang says: don’t ask for ROI, let a thousand flowers bloom.

I say: you already have one—and you’re killing it.

∗ ∗ ∗

The Evidence Is Piling Up

Amazon is not alone in this. A new analysis of 164,000 workers by ActivTrak, reported in the Wall Street Journal, found that AI adoption is increasing the speed, density, and complexity of work rather than reducing it. Time spent on email, messaging, and chat tools more than doubled. Time spent on focused, uninterrupted work—the kind required for solving complex problems, the kind that produces blooms—fell nine percent.

Meanwhile, Anthropic’s own research suggests the gap between what AI can theoretically automate and what it is actually automating is enormous. Even in software and math—the domains where AI is strongest—94% of tasks could theoretically be handled by AI, but only 33% are being automated today. The gap is not technological. It is institutional. The systems, the knowledge, the human judgment required to deploy AI safely—all of that is bloom-dependent. It requires the very people companies are cutting.

Jack Dorsey’s Block cut 4,000 workers last month and tied the decision explicitly to AI. Salesforce’s Benioff cut 4,000 support roles. The C-suite consensus is that AI investment pays for itself through smaller workforces. But the Amazon outage is the first major public crack in that consensus. The system broke not because AI is bad, but because the humans who would have prevented the failure were gone. The wiki was outdated because the person who updated it had been optimized away.

You cannot lean your way to intelligence. You can only bloom your way there.

∗ ∗ ∗

What This Means for the Rest of Us

If you run a factory, a hospital, a logistics operation, a software team—any system where institutional knowledge matters—here is what Amazon just taught you for free:

The intelligence you need is not in the model. It is in the people who know which wiki pages are stale, which procedures have silent exceptions, which machines sound different on Tuesdays. That knowledge cannot be scraped, tokenized, or embedded. It was formed over years of hands-on engagement with a specific system in a specific place. It is, in every meaningful sense, already paid for.

And the ROI case for replacing it? Ask the people who built the spare tire robot whether three-to-five-year payback held up when they finally counted the real costs. Ask Amazon whether the savings from 30,000 layoffs held up when the AI inferred from an outdated wiki and the store went dark for six hours.

The Bloom System asks one diagnostic question: Is your organization blooming—or just surviving?

Amazon, this week, gave us the answer to that question at the largest possible scale. They pulled the bloom out by the roots. And then the AI, left alone in the dirt, inferred from an outdated wiki and broke the store.

∗ ∗ ∗

Dr. Venki Padmanabhan is a plant manager at Advanced Drainage Systems in Wooster, Ohio, where he manages sixty employees across three production lines. He holds a PhD in Industrial Engineering from the University of Pittsburgh and has led manufacturing operations at GM, Chrysler, Mercedes-Benz, Royal Enfield (as COO/CEO), and Ather Energy (as COO). He is the author of the forthcoming book Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away and the founder of the Capability Capital Institute. His Substack, The Long Game for All, is at thelonggameforall.substack.com.

You can watch the author express this essay on YouTube.

Sam Altman has made an admission. Speaking at the India AI Impact Summit last week, the CEO of OpenAI told CNBC-TV18 that companies are engaged in what he called “AI washing” — blaming artificial intelligence for layoffs that would have happened anyway.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.

This is a remarkable thing for the CEO of the world’s most prominent AI company to say out loud. It is also, when you sit with it for a moment, one of the most revealing statements anyone in Silicon Valley has made in years.

Because Altman is not simply observing that companies exaggerate AI’s role in their workforce decisions. He is inadvertently confessing something far more important: that the entire displacement narrative — the story that justifies a two-hundred-billion-dollar industry — is substantially, perhaps primarily, a fiction. And the truth underneath that fiction is one that no technology company wants to confront.

• • •

Watch the needle Altman is trying to thread. On one hand, he needs corporations to believe that OpenAI’s technology is powerful enough to replace expensive human labor and justify enormous licensing fees. The product must be seen as a substitute for human capability — that is what the customers are buying.

On the other hand, he would rather not be blamed for millions of eliminated jobs. “We’ll find new kinds of jobs, as we do with every tech revolution,” he assured the audience in India.

The problem is that both things cannot be true at once. You cannot sell a technology on its ability to replace human labor and then express surprise when companies use it — or claim to use it — to replace human labor. Unless, of course, the automation was never really the point. Unless the companies buying AI were doing something else entirely — something that predates artificial intelligence by decades — and AI simply gave them a more respectable story to tell.

That is exactly what is happening. And the data proves it.

• • •

According to Challenger, Gray & Christmas, approximately 55,000 layoffs in 2025 were attributed directly to AI. Significant — but less than one percent of all job losses for the year. A paper from the National Bureau of Economic Research found that ninety percent of executives said AI has had no impact on workplace employment over the past three years. Nine out of ten people actually making hiring and firing decisions.

Amazon cut 14,000 jobs while telling employees that AI meant the company would “need fewer people.” Six months later, the company walked it back — AI was not actually the reason. Announce layoffs. Cite AI. Wait six months. Admit AI was not the reason. The press release said innovation. The spreadsheet said cost reduction. The AI story was cover.

The freshest example is Block. On March 5, Jack Dorsey laid off approximately 4,000 employees — nearly half the company — and framed it as AI-driven transformation. Then Block’s own former head of communications Aaron Zamost published a response in the New York Times: the roles eliminated were disproportionately in policy and diversity functions. “Standard prioritization and cost management,” he called it, “not an AI-driven reinvention.” He also identified the investor dimension clearly — it matters less whether a company knows how to deploy AI and more whether investors believe it is on track to do so.

Altman calls this AI washing. I want to call it what it actually is: leadership laundering.

• • •

I have spent thirty-six years managing manufacturing operations across three continents. I have been in rooms where automation investments are approved, and on factory floors where the consequences land on actual human beings.

Here is what I can tell you from the plant floor: the vast majority of companies operate their workforce at a fraction of its cognitive capacity. Not because the workers are incapable. Because the organization has never bothered to develop, engage, or deploy the intelligence those workers bring to work every morning.

I call this the false baseline. The assumption embedded in every ROI calculation and automation business case — that the current performance of your workforce represents the full extent of what your workforce can do. It does not. In my experience, most organizations operate their people at thirty to forty percent of cognitive capacity. The remaining sixty to seventy percent goes home every night, unused, uninvited, and eventually unwanted.

This is what leadership laundering looks like in practice. A company spends years treating its workforce as a cost to be minimized. Workers are trained to execute, not to think. Their ideas are not solicited. Their intelligence — which appreciates with every year of experience — is systematically suppressed by management systems designed for control rather than capability.

Then AI arrives. And suddenly the company has a story. We are not eliminating jobs because we failed to develop our people. We are eliminating jobs because technology has made our people obsolete. The fault belongs to the future, not to us.

The worker who was never given a chance to demonstrate her full capability is now described as having been “replaced by AI.” The manager who never asked for her ideas is now a “change agent navigating digital transformation.” The executive who cut the training budget fifteen years ago is now “positioned for the future.”

You cannot claim AI replaced your workers’ contribution if you never measured that contribution. You cannot say a machine made your people redundant if you never made your people essential.

Altman sees the washing. He does not see what is being washed.

• • •

I know what happens when you go the other direction. At Royal Enfield in Chennai, I inherited a workforce the organization had largely written off — operating under low expectations, limited training, minimal engagement. By every automation consultant’s model, they were candidates for replacement.

We did not replace them. We deployed them. The same workforce. The same hands and minds. Different systems. Profits grew twentyfold. Not because we automated. Because we activated.

• • •

But Altman was not finished. Responding to concerns about AI’s enormous energy demands, he offered a comparison that reveals something far more troubling than corporate spin.

“It also takes a lot of energy to train a human,” he told the audience. “It takes, like, 20 years of life and all of the food you eat during that time before you get smart.”

Read that again. The CEO of the most powerful AI company on earth looked at twenty years of a child’s life and saw a training cost.

This is not a gaffe. This is the false baseline rendered as worldview. When you see a human being as something you train — when the food a child eats is an energy input and twenty years of scraped knees and bedtime stories and algebra homework are overhead — you have already made the decision that your technology ratifies. Humans are expensive, slow, inefficient models that happen to run on protein instead of silicon. The twenty years are a sunk cost. And if the machine produces that output faster and cheaper, the human becomes redundant.

I have three children. My eldest is a hematology-oncology fellow who fights cancer for a living. He was not trained to do this. He was formed — by family, by values, by watching his parents navigate three continents of uncertainty, by something in how he was raised that made him believe hard paths were worth walking. No loss function optimized for that. No gradient descent produced the moment he decided the hardest specialty in medicine was the one that mattered most. That was not a training outcome. That was a human being becoming himself.

Altman cannot see the difference. And that is not merely a personal failing. It is a structural one. When your entire business model depends on the premise that human cognition is a commodity — something that can be replicated, scaled, and sold by the token — you must eventually arrive at the conclusion that the original version is just an expensive prototype.

• • •

There is one dimension of this the sharpest critics have not fully reckoned with. The debate about AI and human knowledge has focused on the stock — the vast body of existing human output used to train these models. But the deeper crisis is about the flow. Not the knowledge that has been created, but the knowledge that would have been created if we kept investing in the people who produce it.

Every paper, manual, and insight that trained these models was paid for in human time. In human error. In human life. Now the industry proposes to replace the people who generated that material. In manufacturing, we have a name for what happens when a company eliminates its most experienced workers to cut costs and discovers five years later that nobody can diagnose the problems the veterans used to catch before they became catastrophes. We call it the capability gap. The knowledge lived in the hands and judgment of people who were told they were too expensive to keep.

AI is creating a civilizational capability gap. The models are getting better at reproducing what humans have already thought. They are getting no better at producing what humans have not yet thought — because genuine new knowledge is not a pattern in existing data. It is a break from existing patterns. It is what a forty-year machinist knows in year forty-one that surprises even him. You cannot train a model on knowledge that does not yet exist. And if you eliminate the people who would have produced it, that knowledge will never exist.

• • •

The AI washing will continue. Companies will keep citing artificial intelligence for workforce decisions that have nothing to do with artificial intelligence, because the alternative — admitting they never invested in the people they are now discarding — is an admission no earnings call can survive.

But some of us are watching. Some of us have been on the factory floor at midnight, and in the boardroom the next morning, and at the kitchen table where a boy decides to spend his life fighting cancer — and we know the difference between training a model and raising a human being.

Sam Altman is right that companies are AI washing. He is wrong about what is being washed. It is not the layoffs that need laundering. It is the decades of leadership failure that preceded them. It is the worldview that looks at a child and sees a cost. It is the poverty of imagination that mistakes efficiency for intelligence, output for purpose, and computation for life.

The intelligence was there. It was always there. It was already paid for. Not in kilowatt-hours. In love, in patience, in twenty years of cereal at the kitchen table.

Nobody bothered to unwrap it. That is not a technology story. It is a leadership story — and the leaders are the last ones who want you to know it.

Dr. Venki Padmanabhan is Plant Manager at Advanced Drainage Systems with 36 years of manufacturing leadership experience across three continents. He previously served as COO/CEO at Royal Enfield and COO at Ather Energy. His book, Already Paid For: Why Unlocking Frontline Intelligence Beats Automating Workers Away, is forthcoming. Subscribe to The Long Game at thelonggameforall.substack.com.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.

If you like, you can watch the author express the essay to you on YouTube.

Wally Vinton was the worst team leader in my Trim Shop.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.

That was the consensus, anyway. I was running the Trim Shop at Lansing Grand River Assembly — the section of the plant where Cadillacs got their interiors. Electrical harnesses, carpets, instrument panels, and all the exterior glass installed in a delicate dance between robots applying urethane and humans squeezing windshields and backlites into metal bodies. Hundreds of operations, thousands of potential failure points, and a team of UAW members who had to get every one of them right on a moving line.

Wally was a team leader responsible for harness installation. And for months, he was a disaster. Arguing with supervision. Cussing. Stopping the line. Refusing to complete assignments. Calling the committeeman over every grievance he could manufacture. Management pulled me aside more than once: Why do we keep this guy? He’s killing our numbers.

I understood the frustration. His area was a mess. Downtime was high. His team was demoralized. The easy answer — the one American manufacturing has perfected over decades — was to write him up, document the pattern, and push him out. Replace the problem. Move on.

But something didn’t sit right.

Because Wally wasn’t lazy. Lazy people coast. Lazy people do the minimum and disappear into the herd. Wally was loud. Wally was angry. Wally was stopping the line and demanding that someone pay attention. That’s not the behavior of someone who doesn’t care. That’s the behavior of someone who cares so much it’s coming out sideways.

So instead of writing him up, I went to the floor. Not to lecture. To listen.

And what I heard changed how I think about manufacturing.

The harness installation area was a disaster — but not because of Wally’s team. The harnesses themselves were the problem. They arrived from the supplier taped so thick with protective wrapping that installers were cutting their hands trying to unwrap them. The tape adhesive left residue that made the connectors difficult to seat. Paint from the body shop was filling locator holes that the harness clips needed to snap into. And the rubber grommets that sealed the harness pass-throughs? They were spec’d so tight that pushing them into the floor holes required enough force to cause repetitive strain injuries.

Wally’s team wasn’t failing. They were fighting a war against upstream problems that nobody had fixed because nobody had listened to the people who encountered them eight hours a day, five days a week.

Wally knew every one of these problems. He could describe them with engineering precision. He knew which paint booth was overspray-filling which holes. He knew which harness part numbers had the worst tape adhesive. He knew exactly how many seconds each rubber grommet added to cycle time and how many team members were developing wrist pain.

He had the diagnosis. What he didn’t have — what no one had ever given him — was any reason to believe that sharing it would matter.

Wally had come from other UAW plants. Plants where you learned fast: don’t speak up. Don’t suggest. Don’t stop the line unless you want a target on your back. Plants where the implicit contract between labor and management was simple — we pay you to do what you’re told, not to think. In those environments, intelligence doesn’t disappear. It goes underground. It becomes cynicism, or rage, or quiet disengagement. The system gets exactly what it incentivizes: compliance without contribution. Bodies without brains.

So when Wally arrived at Lansing Grand River, he tested us. His belligerence wasn’t defiance. It was a diagnostic. He was probing whether this plant was any different from the ones that had taught him to stop caring. Every argument, every grievance, every time he stopped the line — he was asking a question he’d never say out loud: Will anyone here actually listen?

What happened next took months. It wasn’t pretty. And it changed everything I believe about manufacturing leadership. I’ll tell you the full story in the essay below.

For months, management’s answer was to see a problem employee. My answer, once I understood what I was looking at, had to be different.

* * *

The turnaround didn’t happen in one conversation. There was no Hollywood moment where I gave an inspiring speech and Wally saw the light. It took months of sustained, unglamorous effort — and it required management to eat crow.

I had to overcome my own distaste for a guy on the floor who was talking back to supervision, cussing out engineers, and making my metrics look bad. That’s hard. Every instinct in a plant manager says: control the disruption. But if I controlled Wally, I’d lose the intelligence.

So we brought him support. Not platitudes. Actual support.

I brought the paint shop supervisor to Wally’s area and made them sit together over the overspray problem. I brought the harness supplier’s design engineer to the floor — not to a conference room, to the actual workstation — and had Wally show him what the tape was doing to his team’s hands. I brought maintenance engineering to look at the grommet tooling and asked them to redesign the insertion fixtures.

Each time, the message was the same: You’re the team leader. You see these problems every day. You have the ability to fix this. We will support you.

Each time, Wally tested whether the support was real. Would the paint shop actually change their booth settings? Would the supplier actually redesign the tape? Would engineering actually build new tooling? Each time, when the answer was yes — slowly, grudgingly, with plenty of setbacks — something shifted.

Wally’s light turned on.

Not all at once. Not dramatically. More like a dimmer switch being turned up over weeks. He stopped filing grievances and started filing engineering change requests. He stopped cussing out supervisors and started pulling them to the floor to show them what he’d found. He stopped stopping the line in protest and started stopping the line to fix root causes.

He organized his team. He built relationships with the paint shop, with harness design, with the grommet supplier. Problem by problem, station by station, they solved it.

Within six months, Wally’s area had the lowest downtime in the entire Trim Shop.

Read that again. The worst team leader — the one management wanted to fire — was running the best area in the shop. Same person. Same paycheck. Same UAW contract. The only thing that changed was that someone finally treated his intelligence as an asset instead of a threat.

And it didn’t stop with Wally. His team members watched what happened. They watched a team leader go from being written off to being listened to. They watched engineering actually show up on the floor. They watched problems that had been ignored for months get solved in weeks. And they started bringing their own observations forward. The guy on Station 12 who’d noticed that a particular clip design failed twice as often in cold weather. The woman on the glass line who’d figured out that the urethane robots were applying adhesive three millimeters off-center on the driver’s side. Intelligence that had been sitting dormant — suppressed by years of being told to just do your job — started surfacing across the whole area.

One Wally became five. Then ten. The Trim Shop didn’t just improve. It transformed.

* * *

Here is what haunts me about the Wally Vinton story, and why I’m writing about it twenty years later.

Wally’s intelligence was always there. On the day he was the “worst team leader in the Trim Shop,” he already had the diagnosis for every problem in his area. He already knew which upstream processes were causing failures. He already had the relationships with his team members to mobilize a fix. All of that capability was present, fully formed, drawing a full paycheck.

And none of it was being used.

Think about this in the language I use when I sit in on venture capital pitches. Wally’s brain — 86 billion neurons, a quadrillion potential synaptic connections, an exaFLOP of processing power running on 20 watts — was being utilized at maybe thirty percent capacity. The thirty percent that follows instructions, shows up on time, and installs harnesses per the standard work. The other seventy percent — the pattern recognition, the systems thinking, the diagnostic capability, the leadership — was stranded. Not because it didn’t exist. Because the system had decided it wasn’t wanted.

We had purchased an exaFLOP processor and were running it at spreadsheet macro level.

When I finally asked Wally what he wanted solved, I wasn’t creating intelligence. I was activating intelligence that was already there. Already paid for. Just waiting for a system that would let it operate.

American manufacturing is full of Wallys. Team leaders and operators and technicians who see problems every day that management can’t see from a conference room. Who carry decades of pattern recognition in their hands and their instincts. Who stopped offering solutions years ago because no one listened, and no one changed anything, and eventually you learn to protect yourself by shutting up.

The cost of that silence is incalculable. Not just in downtime and scrap and quality escapes — though those costs are real and enormous. The deeper cost is that companies are spending billions on automation and AI to replace the very intelligence they suppressed. They’re building inferior robotic substitutes for capabilities that are already standing at the production line, already hired, already trained, already drawing a paycheck.

Already paid for.

I think about Wally every time I see a headline about humanoid robots replacing factory workers. Every time a CEO announces that AI will eliminate thirty percent of their workforce. Every time a consulting firm publishes a study on the ROI of automation.

I want to ask them: Did you try asking your Wallys what they see? Did you bring them the paint shop supervisor? Did you sit on the floor and listen to what the tape adhesive is doing to their hands?

Or did you just decide they were the problem — and start shopping for a replacement?

Wally Vinton is retired now, somewhere in Lansing. I’d love to find him. I’d love to put a microphone in front of him and let him tell this story in his own words — because his version would be better than mine. Angrier, funnier, more profane, and more true.

But even without his voice, the lesson stands. Every plant in America has Wallys. People whose intelligence has been stranded by systems that never asked for it. The most expensive waste in manufacturing isn’t scrap or downtime or warranty. It’s the human capability you’re already paying for and refusing to use.

The light was always on. We just had to stop blocking it.

* * *

Dr. Venki Padmanabhan is Plant Manager at Advanced Drainage Systems and Venture Advisor at Maniv Mobility. He has led manufacturing operations at GM, Chrysler, Mercedes-Benz, Royal Enfield (as COO/CEO), and Ather Energy (as COO) across three continents. He writes The Long Game on Substack.

Thanks for reading The Long Game by Dr. Venki Padmanabhan! Subscribe for free to receive new posts and support my work.