Operating Essay: AI Is Not Killing Jobs. It Is Killing Excuses.
What Keynes, Schumpeter, and 250 Years of Economic History Actually Say About the AI Economy.
Start Here
We are living through a crisis of economics anxiety.
Not an actual economic crisis.
An emotional one, being manufactured at scale.
A viral Substack post imagines AI destroying the economy by 2028. Unemployment at 10%. The S&P down 38%. Markets wobbled. Cable news booked the doomsday panels. The whole episode slotted neatly into a media machine that has learned something profitable: fear converts.
Rage retains.
Victimhood is a growth vertical.
The jobs narrative in 2026 is not being reported. It is being produced.
Every layoff headline stripped of context. Every automation story framed as inevitability. Political narratives woven into employment data not because they illuminate the economic reality, but because they drive eyeballs. Datapoints cherry-picked from whichever time period best serves the thesis the outlet already holds.
Citadel Securities published a rebuttal worth your time. Frank Flight dismantled the scenario with data and economic history. But the deeper point is not that one Substack was wrong. It is that we have an entire information ecosystem optimized to make you afraid.
If you are building something, you cannot let that ecosystem do your thinking.
The economics tell a different story.
Not painless. But knowable. And knowable is enough to build on.
At Operating, we care deeply about what is knowable from the economics.
This is what Operating does every week. I decode the economics of what is actually happening and translate it into frameworks you can build with. Paid subscribers get the full resource library, the Operating Working Group, and every framework I build. $7.99/month.
Keynes and the Infinite Want
In 1930, Keynes predicted that productivity growth would reduce the workweek to fifteen hours by the 2000s. He nailed the productivity forecast.
He missed the labor market entirely.
Because he underestimated human desire. Productivity did not eliminate work. It expanded the consumption frontier. We invented entirely new categories of demand. It is happening again right now. Harvard Business Review reported that in an eight-month study of AI adoption, workers using AI did not work less. They worked faster, took on more tasks, and extended work into more hours of the day.
Keynes’s error is replaying in real time.
This time is once again, not different.
This is Citadel’s core argument. AI is a positive supply shock. It lowers costs, expands output, raises real income, creates demand for things that do not exist yet. Software engineering postings are up 11% year over year. AI adoption at work remains, per the St. Louis Fed, “unexpectedly stable.” The doomsday narrative requires you to believe that this time, for the first time in history, productivity gains will destroy demand rather than reshape it. That is a bet against the entire dataset and against generations of history.
Indeed Job Postings — Software Engineers + Overall Postings, Daily and 21dma. Source: Citadel Securities, Indeed.
Schumpeter’s Gale. The Part Everyone Skips.
Everyone quotes Schumpeter on creative destruction. Nobody finishes the sentence. His point was not that technology destroys. It was that destruction clears ground for creation.
What is growing in the clearing is something Schumpeter could not have imagined. The same AI tools that let a company cut three departments let a first-time founder build what those departments used to do.
The cost of creation has collapsed.
The most important question was never “how many jobs does the technology destroy?”
It was, “what does it make possible?”
Carlota Perez and The Pattern Nobody Wants to Understand
Carlota Perez’s framework maps 250 years of technological revolutions. Every one follows the same arc.
An installation period of speculative capital, hype, bubbles, and fear.
Then a deployment period where institutions adapt and a golden age unfolds.
The dot-com bubble burst in 2000. The internet then spent twenty years becoming the operating system of civilization.
Frenzy first. Golden age second.
We are in the installation period of AI. The panic, the $650 billion in capex, the 2,800 data centers. This is the pattern.
Important Note: The people who build during installation tend to own the deployment phase.
AI Adoption Trends — Share of Working Age Adults Using Generative AI, Real Time Population Survey. Source: Citadel Securities, St. Louis Fed.
The Engels’ Pause: The Honest Part
The Engels’ Pause, coined by Robert Allen, describes the fifty years during Britain’s Industrial Revolution when productivity soared and wages stagnated. It is happening now. Bank of America flagged it: profits gaining ground versus wages, labor income falling as a share of GDP.
But Carl Benedikt Frey identified what broke the pause.
Technology shifted from labor-replacing to labor-augmenting.
The tools of production reached the workers.
Wages caught up.
That shift is happening now, faster than 1840, because the tools are not locked behind capital requirements. They are in your browser. The cure has not been waiting for corporations to share. It has been access to the means of production.
For the first time, that access is virtually free.
What the Economists Say
Keynes: Demand will not collapse. It will reshape.
Schumpeter: Destruction is chapter one.
Perez: This is installation, not apocalypse.
The Engels’ Pause: The transition hurts, but it breaks when the tools reach the workers.
The tools have reached you. And you’re already building with them.
CHART 3: New Business Formation — New Business Applications, US Census Bureau. Source: Citadel Securities, US Census Bureau.
For fifteen years, new business applications in the United States held steady around 200,000 to 250,000. Then in 2020 the line broke upward and never came back down. We are now sustaining 400,000 to 500,000 applications per period.
This is a structural shift. This is millions of people deciding that the economics of building for yourself have changed.
The media is telling you the sky is falling because that is what sells.
The economists and their data are telling you the scaffolding is going up. And the Census Bureau is telling you that people are already climbing it.
Start climbing, friends.
If you haven’t begun, start building.
If you’re already on your construction site, keep going.
Companies are becoming tech stacks.
We are all becoming companies.
- j -
Most people are cosplaying with technology they have opened twice.
84% of humans have never used AI. 0.3% pay for it. Actual adoption is a fraction of what the tools can do. And yet the feed is full of “AI-native founders” who could not build an automated workflow if their revenue depended on it.
The gap between what the technology can do and what people actually do with it is the biggest business opportunity of the next decade.
That is exactly what we work on inside Operating.
Paid subscribers ($7.99/month) get the full resource library, the Operating Working Group, every framework I build, and the complete archive of essays, case studies, and operating toolkits. That is less than two dollars a week to have an economist who actually ran a company decoding the economics of what is happening and translating it into tools you can build with.
Founders ($500/year) get everything above plus two 1:1 (60-minute) strategic sessions with me. Your business model. Your positioning. Your growth problem. We work on it together. This is a second set of eyes from someone who has scaled a company, exited it, and now advises founders building in this economy every day.
The people inside are not posting about AI. They are building their futures with it.
John Brewton documents the history and future of operating companies at Operating by John Brewton. He is a graduate of Harvard University and began his career as a Phd. student in economics at the University of Chicago. After selling his family’s B2B industrial distribution company in 2021, he has been helping business owners, founders and investors optimize their operations ever since. He is the founder of 6A East Partners, a research and advisory firm asking the question: What is the future of companies? He still cringes at his early LinkedIn posts and loves making content each and everyday, despite the protestations of his beloved wife, Fabiola, at times.
Three cheers for the Substack MVPs in replies, the predominant reason many of us newbies joined the platform.
Implicit in John B’s perspective, the train has left the station and won’t be making a U-turn. Leaving us where?
The lively exchange on employment typology, trajectory and productivity is the heart of AI’s societal impact. In reading, my mind went to the very old Zager and Evans tune, “In the Year 2525,” projecting the image of “…legs got nuthin’ to do, some machine doin’ that for you.” When the song sees future “arms hangin’ limp at their sides,” it overlooks the innate, human need for challenge, the nudge of curiosity.
I’d also add that economic prognostication and empirical assessment often leave a gap – quality of life. “Happiness” metrics, such as Bhutan’s GNH are qualitative, maybe not mainstream, yet still valuable in their outlook, and I believe influential in how we gauge AI and humanity’s future.
As innovators innovate, I don’t see masses of “lemmings” awaiting a next directive, but new-found activities that will push more souls to climb Maslow’s model.
Thank you for the Friday morning thoughts, a very pleasant wake-up call.
This post glosses over the distributional impacts of productivity in the 21st century. In the last 20 years, two thirds of productivity gains have come from 20% of the workers (the knowledge economy). Like 25% have come from the software NAICS codes (2-3% of the U.S. workforce). When you de-average everything, the optimist story is not nearly as clean as the headline dataset.
There is a similar argument to consumption. Part of why the frontier of consumption has expanded is because we’ve transactionalized larger and larger swathes of the human experience. You can argue whether that is good or bad - but you can’t argue that it has happened. Or that it’s been a big consumption / GDP tailwind.
Anyway I agree - time to build. But the historical record is not one that paints a picture of abundant opportunity. It’s one of a narrow set of tail outcomes.