Start with a company that has nothing to do with software. Chevron announced plans to eliminate 15% to 20% of its global workforce by the end of 2026, a cut of roughly 8,000 people. The company cited the need to “simplify its organizational structure” and “leverage technology to enhance productivity.”

Chevron is an oil company. It does not write code. It does not sell software subscriptions. Yet it is restructuring at the same scale and with the same language as Oracle and Microsoft.

Walmart restructured its corporate workforce in 2025, eliminating market coordinator roles and layers of regional management. The stated rationale was leaner operations and retail efficiency.

Heineken announced plans to cut up to 6,000 jobs from its global workforce, citing softening consumer spending.

Volkswagen committed to eliminating 50,000 jobs across its Audi, Porsche, and core operations by 2030, because its combustion-era workforce does not map onto its electric-vehicle future.

Dow Chemical announced 4,500 job cuts amid weak industrial demand. Law firm Baker McKenzie is cutting between 600 and 1,000 support staff as it shifts toward AI-assisted research and document review.

None of these companies hired massively during COVID and are now correcting. An energy company, a brewer, an automaker, a chemical giant, and a law firm all restructuring at the same moment points to something more systemic than a hiring hangover.

Four Drivers of the AI Washing Story and the Impact on Workers

The Corporate Fiction and Why It Exists

The dominant narrative in the business press offers two explanations for the current layoff wave. The first is AI displacement: companies are replacing human workers with software.

The second is the pandemic correction: companies over-hired during the COVID boom and are now trimming back to reality. Marc Andreessen argued forcefully on the 20VC podcast that most large companies are overstaffed by at least 25%, with some overstaffed by 75%, and that AI is a convenient cover story for cleaning house.

Both explanations contain truth. Neither explains Chevron, Heineken, or Volkswagen.

Deutsche Bank analysts coined the phrase “AI redundancy washing” to describe companies attributing job cuts to AI when the real drivers are more complicated. The phrase is useful, but it stops one step short of naming what is actually happening.

Companies are not simply lying about their reasons for cutting jobs. They are navigating a genuine, multi-layered business transformation and reaching for the simplest available narrative to explain it to investors, employees, and the public.

That narrative happens to be AI. It also happens to be incomplete.

Subscribe now

The Four Real Drivers

The first driver is skill-set obsolescence, and it is the most underreported story in the current restructuring wave. BCG analyzed 165 million U.S. jobs and found that 50% to 55% will be reshaped by AI over the next two to three years. Reshaped means the work changes, not that the job disappears.

A Chevron operations coordinator whose role was built around managing paper-based compliance workflows is not being replaced by an AI system. They are being made redundant by a Chevron that no longer needs that particular set of skills.

The company is not trimming fat. It is swapping one skill profile for another. The workers losing jobs are not surplus. They are mismatched.

The second driver is investor signaling, and it is arguably the most cynical force in the current environment. Harvard Business Review found that 60% of organizations had already reduced headcount in anticipation of AI delivering productivity gains, while only 2% had made layoffs tied to proven AI implementation.

Companies are cutting based on AI’s projected capability, not its demonstrated performance. When a CEO announces layoffs alongside an AI investment pledge, the stock market rewards both decisions simultaneously.

Workers are losing jobs to a technology that has not yet replaced them, in service of a financial narrative that benefits shareholders before it benefits operations.

The third driver is capital reallocation, and it is the most concrete of the four. Amazon committed $200 billion to AI infrastructure in 2026 while simultaneously announcing 16,000 layoffs. Oracle cut up to 30,000 workers while accelerating its cloud and AI buildout.

The pattern is consistent: labor budgets are being converted into infrastructure budgets. Workers are not being replaced by machines in any direct operational sense. They are being defunded in favor of data centers.

The business is reorienting its spending, and human payroll is the most flexible budget line available.

The fourth driver is organizational flattening, and it is the least visible of the four. JPMorgan Chase has adopted a strategy of redeployment over hiring, using AI to automate back-office coordination while moving existing employees into client-facing roles.

What this describes is a company removing the management and coordination layers that AI can now handle. A company that needed five layers of management to move information through its structure may genuinely need three if AI handles the middle.

The workers losing jobs in that scenario are not being replaced by AI. They are being displaced by a flatter organizational chart that AI makes possible.

That is a structural change, not a workforce reduction.

What This Means for You

The gap between the corporate narrative and the operational reality matters enormously for anyone making a career decision right now.

If you believe the AI displacement story entirely, you retrain for AI-adjacent skills and hope to land in a role that survives automation. If you believe the pandemic correction story entirely, you wait for the labor market to stabilize and assume your sector will recover. Both responses are wrong, because both stories are incomplete.

The more accurate picture demands a harder look. The IMF found that in regions with higher demand for AI-related skills, employment levels in AI-vulnerable occupations are 3.6% lower after five years, even accounting for workers who acquired those skills.

Learning AI tools raises your wages. It does not necessarily protect your field.

The skills you build matter less than the sector you build them in and the organizational structure your employer is moving toward.

The workers most at risk right now are not the workers AI can replace most easily. They are the workers whose roles exist primarily to coordinate, translate, or buffer information inside large organizations.

Middle managers, administrative coordinators, compliance processors, regional supervisors: these are the roles that organizational flattening targets first. AI does not do their jobs. It makes their jobs unnecessary by connecting the layers above and below them directly.

The Dallas Fed research established that AI automates the codifiable tasks entry-level workers perform while making experienced workers more valuable. That pattern holds across all four of the drivers we looked at.

Experience, judgment, and tacit knowledge are the durable assets. Titles, coordination functions, and process management roles built around workflows that AI can now handle are the vulnerable ones.

The practical translation for a middle-American service worker is this: the question worth asking about your job is not whether AI can do it. The question is whether your employer needs the organizational layer your job occupies.

Those are different questions.

The first is about technology. The second is about strategy. Right now, corporate strategy is moving faster than the technology, and workers are bearing the cost of the gap between them.

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now

The Fault Line Nobody Is Disputing

This past week, four independent research organizations, using four different data sources, reached the same conclusion: The labor market is splitting.

Workers with experience and AI fluency are pulling ahead. Workers without either are falling behind.

Goldman Sachs, analyzing payroll data, found that the unemployment and wage gap between workers under 30 and workers aged 31 to 50 has widened sharply.

Further, the Dallas Fed, examining wage premiums across 205 occupations, found that AI tends to automate the codifiable tasks that entry-level workers depend on while making experienced workers more valuable.

Meanwhile, the IMF, drawing on job posting data across advanced economies, found that entry-level hiring is declining in AI-exposed fields even as wages rise for workers who hold those jobs.

And PwC’s Global AI Jobs Barometer found that workers with advanced AI skills earn 56% more than peers in identical roles without those skills.

The corrobaration of findings is as close to settled as anything in this still-early field.

Share

The Reshaping Frame Is Winning the Research Debate

The apocalyptic displacement narrative lost ground this week to a more precise and arguably more useful frame. BCG analyzed 165 million U.S. jobs across 1,500 roles and found that 50% to 55% will be reshaped by AI over the next two to three years.

Reshaped means workers keep their roles but face fundamentally different expectations. MIT’s Computer Science and Artificial Intelligence Laboratory tested 11,500 real-world tasks against more than 40 AI models and concluded that AI is advancing like a rising tide rather than a crashing wave.

SHRM’s State of AI in HR 2026 report found that AI is 5.7 times more likely to shift job responsibilities than to eliminate jobs outright.

The practical meaning for a service worker is this: the more realistic near-term threat is not losing your job to AI. It is being required to do your job differently, faster, and with AI tools you may not yet know how to use, while your employer adjusts expectations upward without adjusting pay to match.

Anxiety Is Running Ahead of Actual Worker Displacement

Mercer’s Global Talent Trends 2026 report found that worker fears about AI job displacement jumped from 28% in 2024 to 40% in 2026.

The Ipsos survey conducted for Google across 21 countries found that Americans are split exactly 50-50 on whether AI will ultimately help or hurt workers.

Nearly three in four Americans believe the government should act to prevent AI-driven job loss.

Yet, the Epoch AI and Ipsos survey published Thursday found that only 20% of full-time workers say AI has actually taken over parts of their job.

The gap between anxiety and documented worker displacement by AI is real and wide. That does not mean the anxiety is wrong.

It may mean the displacement is coming and workers sense it before the data fully captures it. It may also mean the headlines are running ahead of the evidence. Both possibilities deserve serious attention.

Short-Term Trends (Now to 12 Months)

The most concrete short-term finding of the week came from Challenger, Gray & Christmas, the outplacement firm that tracks corporate layoff announcements. The group reported that AI was cited as the leading reason for tech layoffs in March 2026, accounting for 25% of stated reasons. Just one month earlier, that figure was 10%.

The AI-attributed share of layoffs more than doubled in a single month. The speed of that acceleration matters more than the absolute number.

Running alongside the announced layoffs is a quieter restructuring that the Novoresume survey of 1,000 U.S. workers documented in detail. Nearly half of AI-using workers have already compressed their full day’s output into fewer hours using AI tools. They spend the remaining on-clock time on personal activities without their employer’s knowledge.

The work is still getting done. The employer does not yet know the workday has effectively shortened.

When employers do figure this out, the response could go several ways: demanding more output for the same hours, cutting hours and pay, or restructuring roles entirely.

That reckoning has not yet arrived for most workers. The data suggests it is approaching.

Meanwhile, physical AI deployment in manufacturing is a parallel story that white-collar coverage has largely missed. A Deloitte survey of more than 3,200 global business leaders found that 58% are already using robotic systems guided by machine learning in their operations. That number rises to 80% when executives describe their plans for the next two years.

Carmakers. including Audi and BMW, are already piloting humanoid robots in production settings. Oxford Economics projects that global manufacturing could lose up to 20 million jobs by 2030 as this wave accelerates.

This worker displacement through automation runs on a different timeline from the generative AI wave hitting office workers: It hits communities that have already absorbed decades of offshoring and earlier rounds of automation. This second wave for manufacturing is landing on already weakened ground.

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now

An Audit of How 1000 Workers Use AI

, Co-Founder & Career Expert over at Novoresume, messaged me about how a significant number of job seekers are using AI. His company just published a survey of 1,000 US workers on what they’re actually doing with AI at work in 2026.

He told me, “One finding that surprised me is that 1 in 7 workers used AI to get their current job by having it whisper answers during a live video interview. And 6 in 10 feel zero guilt about it.”

Share

Also from the survey, when workers were asked if they could perform their jobs at the same level without AI, 68% said yes absolutely, it is just a convenience. Another 28% said probably, but their work would be slower. Only 4% said they would struggle significantly or could not do their job at all.

The finding implies workers are not becoming dependent on AI.

So what would workers actually do if AI were banned?

Fifty-three percent said it would not affect them much. On the other hand, 13% said they would welcome a ban. Only 7% said they would quit or start job hunting. Ten percent admitted they would keep using AI secretly.

What about worker feeligs about job security? When asked whether AI will eventually make their job obsolete, 58% of AI-using workers said their job is safe from AI entirely.

Of those who do see obsolescence on the horizon, only 5% believe it is coming within two years.

Workers who use AI daily seem, on the whole, less worried about displacement than the research community studying them.

A Major New Survey: One in Five Workers Says AI Has Taken Over Part of Their Job

The biggest story breaking this morning comes from a survey released today by Epoch AI, a nonprofit research organization, conducted in partnership with Ipsos. The poll of 2,000 American adults found that half of all American adults used AI in the past week.

Among full-time workers, Epoch AI wrote, 20% said AI has taken over parts of their job.

The replacement figure outpaces the creation figure: only 15% of full-time workers said they had begun doing new tasks that AI made possible. That five-point gap between replacement and creation is the number worth watching.

Nicholas Miailhe, an AI policy expert at the Global Partnership on Artificial Intelligence, told NBC News: “When 1 in 5 workers say AI is already replacing parts of their job, we can start talking about labor market restructuring happening in real time. The fact that replacement seems to be outpacing augmentation should draw our attention: the policy window to shape how AI transforms work is probably closing faster than most governments realize.”

The survey also found that among workers who used AI in the past week, roughly half used AI between two and five days per week. The most common AI tasks were looking up information or recommendations (80%), writing or editing text (59%), and brainstorming ideas (53%).

ChatGPT led all AI services at 31%, followed by Google’s Gemini at 21% and Microsoft’s Copilot at 10.5%. One detail worth noting: roughly half of American adults who used AI for work in the past week used their own personal subscriptions rather than tools provided by their employer.

So in addition to IT departments having to contend with BYOD (Bring Your Own Device) policies and infrastructure, now they have to deal with BYOAI (Bring Your Own AI) issues.

MIT: A “Rising Tide,” Not a “Crashing Wave”

A study from MIT’s Computer Science and Artificial Intelligence Laboratory, covered this week by Axios, pushes back on the most apocalyptic job-displacement predictions.

The research tested more than 11,500 real-world tasks drawn from the U.S. Labor Department’s database against more than 40 AI models. It found that AI advances across the workforce more like a “rising tide” than a “crashing wave” That means change will be broad and gradual rather than sudden and concentrated.

The variation in AI success rates by sector is instructive. AI achieves its lowest success rate in legal work at 47%, where precision and strategic judgment remain essential. It achieves its highest rate in installation, maintenance, and repair at 73%. The primary use of AI in contexts is to automate administrative and documentation tasks within those trades.

In management, AI achieves a 53% success rate on planning, writing, and analysis tasks. It falls short on coordination, judgment, and decision-making.

The MIT finding is that we are several years from AI achieving near-perfect success rates across most occupations. Workers may have more time to adapt than the loudest headlines suggest.

The IMF’s Skill Gap Warning: A Middle-Class Problem

A new IMF Staff Discussion Note titled “Bridging Skill Gaps for the Future: New Jobs Creation in the AI Age” contains a finding that deserves more attention than it has received.

About one in ten job vacancies in advanced economies now requires at least one new AI-related skill. The United States leads this trend globally, with emerging economies seeing roughly half that rate.

These new skills pay more: job postings requiring four or more new skills pay up to 15% more in the UK and 8.5% more in the United States. But the distributional picture is troubling.

High-skill and low-skill workers tend to gain the most from new-skill adoption. Middle-skill workers in routine office jobs are getting squeezed out.

The IMF adds a finding that complicates the “just learn AI skills” advice: in regions with higher demand for AI-related skills, employment levels in AI-vulnerable occupations are 3.6% lower after five years than in regions with less AI-skill demand. Acquiring AI skills raises your wages. It does not necessarily protect the overall employment level in your field.

SHRM: AI Changes Job Responsibilities More Than It Eliminates Jobs

The Society for Human Resource Management (SHRM) released its State of AI in HR 2026 report this week. The data finds that AI’s organizational impact is 5.7 times more likely to shift job responsibilities than to displace jobs outright. It is three times more likely to create new roles than to eliminate existing ones.

That framing aligns with the BCG and MIT findings we reported yesterday and earlier this week.

The SHRM data also reveals an adoption gap by seniority: by 2025, 73% of HR professionals at the director level and above had adopted AI for work purposes, compared to 65% of individual contributors.

Leadership is driving AI adoption. Frontline workers are catching up. However, the gap between senior and entry-level AI fluency is real and measurable.

The Ipsos Sentiment Divide: Americans Split Exactly Down the Middle

A separate Ipsos survey on AI attitudes, conducted across 21 countries for Google, contains one finding that captures the current public mood precisely.

Respondents are split exactly 50-50 on whether AI in the workplace will ultimately create jobs and help workers, or eliminate jobs and hurt workers.

A separate Ipsos data point from January adds important context: nearly three in four Americans think the government should take action to prevent job loss from AI. More than half think increased AI use will lead to more inequality and polarization, a nine-point increase from 2023.

Public concern is rising faster than public understanding.

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now

Axios and Morgan Stanley: The Picture Is More Complicated

Morgan Stanley frames the AI jobpocalypse situation as two-sided: the same technology that automates tasks can also raise productivity and support hiring.

The radiologist example is instructive here: A decade ago, AI pioneer Geoffrey Hinton predicted deep learning would make radiologists obsolete within five years. The opposite happened: radiologists adopted AI as a tool, their numbers increased, and their pay went up.

The harder-to-dismiss finding from Morgan Stanley is that companies are talking about AI displacing workers far more often on earnings calls than they are talking about AI driving new hiring. That gap between rhetoric and reality is a signal worth tracking.

Share

Software Engineering: A Counterintuitive Data Point

Against the grain of the software engineering displacement narrative, data from TrueUp, a tech hiring analytics firm tracking more than 260,000 job openings across 9,000 companies, shows that demand for software engineers has rebounded sharply to levels not seen in over three years.

Software engineering roles have roughly doubled since their low point in mid-2023, with more than 67,000 positions currently open. The explanation is actually pretty straightforward: companies investing heavily in AI need skilled engineers to build and maintain these systems. The AI buildout is creating demand even as it displaces workers in other categories.

Fortune’s “9 Reasons” Counterargument

Fortune magazine’s April/May issue features a piece arguing that the jobpocalypse panic is premature. Many layoffs attributed to AI are driven instead by financial underperformance or earlier overhiring, with AI serving as a convenient cover story.

The piece cites what it calls the Klarna Effect: in early 2024, the Swedish fintech company announced that AI was doing the work of 700 customer-service employees. By spring 2025, Klarna had reversed course and was hiring again, having decided that human workers were still necessary for key functions.

China: A Different Story, For Now

CNBC’s China Connection newsletter, published today from Beijing, reports that AI-driven layoffs at Chinese companies look far more contained than in the United States. Two factors explain the difference.

The Chinese government actively directs companies to retain workers rather than automate them out of jobs.

And labor costs are dramatically lower: the average monthly salary for high-demand algorithm engineers in China is about $2,900, compared to the $25,000-per-month compensation that a comparable U.S. engineer might earn in Silicon Valley. When labor is cheaper, the economic case for replacing workers with AI is weaker.

Europe: AI-Intensive Firms Are Hiring More, Not Less

The European Central Bank published research this month drawing on a survey of 5,000 European firms.

Companies that make significant use of AI are about 4% more likely to hire additional staff than their non-AI peers, and firms that invest in AI are nearly 2% more likely to hire.

The ECB’s explanation: firms using AI for research, development, and innovation tend to grow, which creates demand for new workers.

The employment effect is driven primarily by small firms, while AI appears neutral for large firms’ employment decisions.

Separately, a report cited today warns that up to 200,000 jobs could be eliminated in European banking alone as banks accelerate generative AI adoption in customer service, back-office operations, and basic financial analysis.

The Global Corporate Layoff Ledger

A running tally from Programs.com, updated this week, puts the cumulative count of explicitly AI-attributed corporate layoffs since 2023 at over 160,000 workers across major employers.

Citigroup has announced plans to cut roughly 20,000 positions through AI-enabled automation of middle-office and operational functions. That cut has not yet fully materialized, but it is on the books. Across all industries, more than 45 CEOs have publicly announced layoffs they attribute directly to AI efficiency gains.

So Far This Week

AI’s impact on the labor market is accelerating but deeply uneven. Goldman Sachs estimates a net loss of 16,000 U.S. jobs per month, with AI substitution eliminating roughly 25,000 roles monthly while productivity gains add back only 9,000.

The tech sector shed 52,000 jobs in Q1 2026, a 40% year-over-year jump, with AI cited as the leading reason for March layoffs. Entry-level workers and recent graduates are absorbing the hardest hits, and four decades of Goldman data show that technology-displaced workers carry a lasting earnings scar for up to ten years after job loss. Yet counterforces are real: software engineering hiring is rebounding sharply,

European firms that invest heavily in AI are hiring more workers rather than fewer, and China’s government is actively restraining automation-driven layoffs. Deutsche Bank has coined the term “AI redundancy washing” to describe companies using AI as cover for corrections to pandemic-era overhiring.

The core question remains unresolved: how much of the disruption is genuine displacement, and how much is corporate storytelling?

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now

The Number That Matters: 16,000 Jobs Lost Per Month

The biggest piece of new research, published yesterday and picked up by news outlets everywhere, comes from Goldman Sachs, one of Wall Street’s most influential banks. Goldman finds that AI is producing a net loss of about 16,000 jobs per month over the past year. While automation eliminated roughly 25,000 roles monthly, only about 9,000 jobs were added through productivity gains.

To put that in perspective: the U.S. economy has about 160 million workers. So 16,000 a month is a small fraction right now. But the direction of travel matters more than the current size of the number. It is moving in one direction, and it is accelerating.

Share

Who Is Getting Hit the Hardest: Young Workers

The Goldman research also identifies who is absorbing most of the pain, and the answer is younger workers. Entry-level employees, particularly those under 30, are overrepresented in roles most exposed to AI substitution. As a result, unemployment and wage gaps between younger and more experienced workers have widened.

A second Goldman report, published Monday, looked at what job displacement actually costs workers over the long run. The findings are sobering. Tracking more than 20,000 individuals across four decades of data, Goldman economists found that workers whose jobs are eliminated by technology don’t just struggle in the short term.

Over the ten years following a job loss, real earnings for technology-displaced workers grew nearly 10 percentage points less than for workers who were never displaced.

That is a long financial scar to carry. Lose your job to automation at 28, and you may still be feeling it financially at 38.

The twist, though, is that Gen Z may actually prove more resilient than older workers. Younger workers are more likely to switch to new kinds of work and pick up new skills. This may give them an advantage over, say, a 52-year-old whose entire career was built around one set of tasks that AI just learned to do.

A Slower Story Than the Headlines Suggest

Axios published a synthesis this morning pulling together research from Goldman and Morgan Stanley, and the tone is more cautious than the alarming headlines might suggest. Morgan Stanley points out that AI’s impact on labor demand cuts both ways: the same technology that can automate tasks can also make workers more productive, which can actually support more hiring.

The radiologist example is useful here. Ten years ago, a famous AI researcher predicted that deep learning would make radiologists obsolete within five years. Instead, radiologists adopted AI as a tool, the number of radiologists increased, and their pay went up. The technology augmented the profession rather than eliminating it.

That said, Morgan Stanley’s analysis of company earnings calls finds that companies are talking about AI displacing workers far more than they are talking about AI driving new hiring. There is a notable gap between what executives say publicly and what shows up in the actual employment data so far.

The CEO of Pearson Says Slow Down

Fortune published an op-ed yesterday from the CEO of Pearson, Omar Abbos. Pearson is a large education company. He pushed back on what he calls the “AI job apocalypse” narrative. His argument is that the panic is mostly a Silicon Valley story, not an economy-wide one.

He cites Oxford Economics research finding the evidence of a broad AI-driven jobs shakeup to be patchy at best. He points out that the U.S. unemployment rate is currently 4.4% overall and 9.4% for workers aged 16 to 24. The stats are actually below the peaks Europe saw during the economic troubles of the 1990s.

His point is that the sky, as of today, has not fallen. What he acknowledges is that specific sectors, particularly software and knowledge work, are already being reshaped in ways that are visible and real.

Microsoft’s AI Chief: 18 Months Until Your Office Job Is Automated

Not everyone is taking the cautious view. Fortune also surfaced a striking prediction this week from Mustafa Suleyman, the CEO of Microsoft AI. In a recent conversation with the Financial Times, Suleyman said he expects AI to reach human-level performance on most professional tasks within 18 months.

He specifically named accounting, legal work, marketing, and project management as vulnerable. He argued that most tasks involving “sitting down at a computer” will be fully automated within that window.

The article notes that the data on the ground does not yet support that timeline. A 2025 Thomson Reuters report found that lawyers and accountants are experimenting with AI for targeted tasks like document review. The results ahow marginal productivity improvements, but well short of mass job displacement.

In some cases, AI has had the reverse effect: one study found that AI actually made software developers’ tasks take 20% longer.

Why Young Workers Specifically Are in the Crosshairs

The Federal Reserve Bank of Dallas published research this month that explains, in clear terms, why entry-level workers are bearing the brunt of company plans for AI. The key is the difference between two types of knowledge.

The first type is knowledge you can look up in a textbook, things like legal procedures, financial formulas, or coding syntax. The second type is knowledge that comes from years of actually doing a job: the judgment calls, the pattern recognition, the feel for when something is about to go wrong.

AI can handle the textbook kind of knowledge well. It has a harder time replicating the experiential kind. So AI tends to automate entry-level roles. Junior staff rely more on the former kind of knowledge. Meanwhile, AI use actually makes experienced workers more valuable, since their tacit knowledge becomes the irreplaceable part of the job.

The wage data backs this up. Since fall 2022, wages in computer systems design have risen 16.7%, compared to 7.5% nationally. Experienced workers in AI-exposed fields are getting paid more. Recent graduates trying to break into those same fields are finding fewer doors open.

The Bottom Line

The Peterson Institute for International Economics published a useful framing piece last month noting that this entire area of research is still, as they put it, “in the first inning.” Studies on AI and productivity find mostly positive effects, with notable exceptions: one found AI made developers slower, another found consultants over-relied on AI for tasks beyond its capabilities and produced worse results.

AI exposure has also been found to increase work hours and decrease employee satisfaction in some studies.

So, as of today:

• the data is real but modest

• the youth employment squeeze is the most consistent finding across multiple research groups

• the “scarring” effects of displacement look serious for those it does hit

• the 16,000 net jobs per month figure from Goldman is probably the number that will be cited most in the coming weeks.

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now

In writing deep dives about AI, AI Agents, and Superintelligent AI (also known as (Artificial General Intelligence —AGI) these past few months, I’ve observed a high level of anxiety workers — employed and otherwise, across a wide range of demographics — are experiencing high levels of anxiety about the impact of AI on their jobs and the job market broadly.

Interest in the area has even shown up in the engagement levels of the articles I’ve published about AI.

There is also a high degree of nervousness about the training and employment prospects for the generation gearing up for college studies and work opportunities that may be available to entry-level talent. We’re talking about a relatively short-term window here, of three- to five-years, during which we will see major disruptions to the labor market.

So I’ve decided to begin putting out daily roundups of news specific to AI and the reshaping of the nature of work, jobs specifically, and Labor market trends. Each day I’ll be posting snippets of news, reports, and statistics relevant to AI and our livelihoods.

The Numbers Are In, and They Are Striking

The most significant news coming out last week was from Challenger, Gray & Christmas, the outplacement firm that tracks layoff announcements. The U.S. tech sector saw 52,050 job cuts in the first quarter of 2026, a 40% jump from the same period a year earlier.

Share

In March alone, AI was the leading stated reason for tech layoffs, accounting for 15,341 of the firings, roughly 25% of the total. Just a month earlier, that figure was 10%. The AI-attributed share of layoffs more than doubled in a single month.

However, attribution is not causality.

In early March, Block CEO Jack Dorsey announced the elimination of 4,000 jobs, roughly 40% of the company’s global workforce. He, in contrast to most other American CEOs, the growing capability of AI tools to perform a wider range of tasks.

Then, right at the end of the quarter, Oracle laid off thousands of employees as part of its push into AI spending, with a brief 6 AM email sent to workers across the US, Canada, Mexico, India, and other countries. Estimates put that Oracle cut somewhere between 20,000 and 30,000 people.

But Is AI Actually the Cause?

Marc Andreessen argued this week on the “20VC” podcast that AI is a “silver-bullet excuse” companies use to justify cuts driven by pandemic-era overstaffing. “Essentially every large company is overstaffed,” he said, most by at least 25% and some by as much as 75%.

After all, stock prices of companies rise when they attach the label “AI” to the layoffs, whereas saying something like “we overhired during COVID” makes CEOs look unimaginative and part of the tech herd as opposed to being real leaders.

Deutsche Bank analysts coined the phrase “AI redundancy washing” to describe companies attributing job cuts to AI. They warned that the phenomenon “will be a significant feature of 2026.”

The CEO of Randstad, the world’s largest staffing firm, told CNBC that “those 50,000 job losses are not driven by AI, but are just driven by the general uncertainty in the market.”

Duke University and the Federal Reserve Banks of Atlanta and Richmond recently released a survey of roughly 750 senior finance executives. Finance leaders said that in 2025, AI’s impact on their firms was negligible, though they expect it to tick up in 2026.

A lot of that has do with the fact that Generative AI implementations are way different from typical tech applications, from spreadsheets to ERP. I wrote about this in my Substack article “Boards Don’t Hit Back".

When weighted to reflect the broader economy, the survey suggests AI could drag down U.S. employment by about 0.4%, or around 500,000 jobs, through a combination of cuts and reduced hiring.

Leave a comment

A compilation of the Substack articles examining how the invasion already happened. You just weren’t invited. $9.95 flat fee for the bundle (PDF, ePUB), no subscription required. 2-hr reading time.

Buy it Now