The fear is the product, people!
I want to tell you about a magic trick. It is one of the oldest tricks in the capital markets playbook, and it has been performed so many times that you would think audiences would recognize it by now. But alas, they do not. They keep gasping at the right moment, every single time.
The trick works like this.
You stand in front of a crowd and you tell them something enormous is coming. Something transforming and civilization-altering that will remake the economy so fundamentally that the old rules no longer apply and old-fashioned concepts such as basic unit economics definitely no longer apply. And you say this with the specific cadence of a man who has seen the future and is generously sharing it with the rest of us before it runs us over. And then, the crowd leans in and the investors write checks, you see the valuation climbs and you, being the instigator, you borrow against the valuation to build more of the thing. And you repeat this trick until wealthy or indicted, whatever comes first.
Sam Altman, Dario Amodei and a bunch of other Big Tech fear mongers are performing this trick in real time, and I have to hand it to them, the production values are extraordinary.
Altman told us beginning of last year that AI agents will join the workforce in 2025, that customer support jobs will be the first to go and that he is personally uncertain about the future of software engineering as a profession. Mister Amodei has told us that 50% of entry-level white-collar jobs could be gone within five years, that unemployment will spike to somewhere between 10 and 20%, and that most CEOs want to use AI to employ fewer people and are “just not aware that this is about to happen yet”. He said this last part privately to other CEOs, which is how we know he believed it, because nobody performs job apocalypse theater for a private audience.
Anthropic’s 2024 revenue was approximately $918 million. Anthropic’s valuation was $183 billion. That is a revenue-to-valuation ratio of 0.5%. OpenAI had revenues of $3.7 billion against a valuation of $157 billion in 2024. Deutsche Bank projects OpenAI’s cumulative losses before profitability could reach $143 billion between 2024 and 2029. This is before you count the $1.4 trillion data center commitment that OpenAI’s CFO announced with the energy of someone announcing they have pre-ordered a very nice lunch.
And let me tell you something new. This fear of people loosing their jobs is not a side effect of the AI revolution.
The fear IS the revenue model.
More rants after the messages:
- Connect with me on Linkedin 🙏
- Subscribe to TechTonic Shifts to get your daily dose of tech 📰
- Please comment, like or clap the article. Whatever you fancy.
What the numbers actually say, since nobody seems to want to mention them
Here is the empirical situation with AI and jobs, assembled from sources that are not a tech CEO on a podcast.
According to the Information Technology and Innovation Foundation in 2024, AI created approximately 119,900 jobs in the United States and eliminated approximately 12,700. That is a 10-to-1 creation-to-destruction ratio. The 12,700 job losses attributed to AI represented roughly 0.1% of all US layoffs that year. The International Labour Organization, which employs actual economists whose compensation does not depend on making AI sound apocalyptic, estimates that at most 2.3% of jobs globally have the potential to be fully automated by AI. MIT economist Daron Acemoglu, a Nobel laureate whose opinions are not underwritten by NVIDIA, estimates that 5% of jobs will be genuinely displaced or significantly transformed over the next decade.
Altman is predicting 40%. Acemoglu is predicting 5% over ten years. These are not slightly different views though. These guys are on different planets.
The Challenger, Gray & Christmas data for January 2026 which I wrote about earlier, tracks the actual stated reasons for layoffs, and they found that AI was explicitly cited in approximately 7% of layoff plans. The leading causes they mentioned were contract losses, market conditions and restructuring. But the thing is that the companies who were announcing those restructuring plans include many of the same companies whose CEOs spend their media appearances explaining that AI is fundamentally transforming the labor market. You see where I’m going at? They are not wrong that the labor market is experiencing turbulence, that’s quite obvious, but they are wrong about the cause.
The cause, in the words of Amazon CEO Andy Jassy when asked about his own company’s cuts, was “over-hiring and too many management layers”. For instance, Microsoft’s headcount remained (in Satya Nadella’s characterization) “essentially flat” through the AI transformation period. IBM announced in May 2023 that it was pausing hiring for 7,800 roles that AI could replace and the thing is they’re now tripling its entry-level hiring in 2026 for the very roles AI was supposedly about to eliminate (sic!). They didn’t offer an explanation by the way. I imagine none was considered necessary because it is not in their interest.
What is in their interest, is that the AI hype continues to generate wealth for them.
Your layoff was not caused by the robot
The specific narrative distortion happening here has a name. Researchers call it AI-washing, which is the practice of attributing workforce reductions to AI when the real drivers are pandemic-era overhiring, margin pressure, slowing consumer demand, or plain old corporate restructuring that nobody wants to explain to a journalist on a Tuesday.
AI-washing works for everyone involved. For the company, it repositions a painful and embarrassing correction as a forward-thinking strategic pivot. They are changing the narrative from “we hired 40% too many people in 2021 because interest rates were zero and we were insane” to “we are boldly embracing the AI future” because that sounds better in press releases. Laying off people because you misread demand is a management failure, but laying off people because AI is transforming your industry is visionary leadership. Same people but completely different headline.
And AI-washing is even better for investors because it confirms the thesis that the model is working so well that human jobs are disappearing in real time, that the disruption is happening, right here, right now, in this earnings call. Please continue to price our equity as if we are printing money, which we are not, but we do have very impressive compute infrastructure commitments and if you squint at them in a certain light they look almost like revenue.
Every layoff that gets attributed to AI is free marketing for the AI companies specifically, and a trembling LinkedIn post from a recently displaced tech worker is a testimonial to the system’s power.
The fear propagates.
Allow me to auto-circle-jerk-reason for a moment. The fear of job losses through AI attracts investment in AI companies, which in turn funds their ever growing demand for computing power, and the compute runs the model that generates the demos, and the demos generate the fear.
This a perpetual motion machine built entirely upon our collective anxiety.
It’s the same ‘fear-mongering’ we see happening in Europe as well, with Rutte’s ‘fear the Russians, pay more taxes’ rant.
The bill that is coming due
Let me talk about what is actually being spent here, because the numbers are the kind that require you to sit down and possibly have a stiff drink.
Take Project Stargate, Sam’s love baby with El Presidente that was announced in January 2025 that commits $500 billion to AI infrastructure by 2029. Meta has committed $600 billion to AI infrastructure. Apple has committed $600 billion as well, and NVIDIA has committed $500 billion and to top it all up, big blue IBM has committed $150 billion and I think I may have forgotten Oracle and a few others, but never mind.
Now, add up the publicly announced infrastructure commitments across major tech companies and you arrive at something north of $3 trillion.
Three trillion dollars is the GDP of the United Kingdom. It is roughly 11% of US GDP. And it is an absurd amount of money that requires us to build a new relationship with the concept of “large amounts of money”.
The energy consumption picture is similarly confrontational. Traditional server racks consume between 5 and 15 kilowatts and AI-optimized racks consume 40 to 60 kilowatts and the most advanced AI training infrastructure hits over 100 kilowatts per rack. McKinsey – as always, the universal number cruncher – projects that AI data centers could consume 11 to 12% of total US electricity by 2030, and at current, US data center consumption is 176 terawatt hours annually but it is projected to reach somewhere between 325 and 580 terawatt hours by 2030. For reference, the entire country of Poland uses about 170 terawatt hours per year.
We are about to add Poland’s entire electricity consumption to power the infrastructure that is supposedly about to eliminate your job.
And the joke is that none of them generates profit at the current moment.
Inhale . . .
OpenAI is selling dollars for $0.70, to borrow a characterization from a Deutsche Bank analyst who was not feeling generous. Anthropic won’t break even until 2028 – at the earliest. The unit economics of running a large language model inference at scale do not currently support the business model, and the expectation that inference costs will fall fast enough to fix this runs directly into the reality that demand and model complexity are rising faster than costs are falling.
The only thing that makes continued investment rational is “the narrative” of AI replacing humans.
The narrative must therefore be maintained at all costs or else the balloons will pop, and the most powerful narrative available is the one that convinces you the technology is already so transformative that it is restructuring the global labor market in real time.
It does not matter that the productivity statistics do not show it nor that that 90% of CEOs surveyed by the National Bureau of Economic Research say AI has had no impact on their employment or productivity over the last three years. Their story must be bigger than the data, because the data cannot justify the valuation, and the valuation cannot be allowed to fall, because the valuation is the collateral for the next round of borrowing.
Exhale . . .
Now for the finale.
Who is actually getting disrupted and why they are not talking about it
This part of ‘the narrative™’ ©® gets less coverage because it is less frightening and therefore less useful to everyone selling something.
Allow me . . .
The actual disruption happening in the economy right now is not in the labor market, but in software industry economics. Traditional (SaaS) software vendors built businesses on per-seat licensing, high implementation costs, extensive support contracts, and the beautiful stickiness of enterprise software that is too painful to replace. And this model is being compressed by AI tools that are somewhat cheaper, faster to deploy, require less maintenance, and increasingly do not require users to learn anything beyond natural language.
In February 2026, the software sector had its worst single day since April of the prior year. Thomson Reuters fell 15.83%, its biggest single-day drop on record. LegalZoom fell nearly 20%. RELX, which owns LexisNexis, fell 14%. SAP fell 16%. ServiceNow fell 11% despite beating earnings consensus. Approximately $1 trillion in market value was wiped out in a few days, not from labor market fears but from investor recognition that the per-seat licensing model is being replaced by consumption-based pricing on AI tools that do not care how many users you have.
This is the real disruption people, and it is happening to Salesforce, Adobe, Atlassian, HubSpot, Intuit, and a hundreds other companies whose products you have probably used today. It is a genuine structural shift in how enterprise software gets priced and sold.
It is certainly interesting and significant and it would make for excellent economic journalism.
But it does not make you scared for your own job, so it does not generate the kind of engagement that a Sam Altman prediction about 40% workforce replacement generates. And engagement, in the current attention economy, is the substrate on which venture valuations grow.
Agree?
Then let’s continue.
The career ladder problem nobody wants to own
To be fair to the data, because I said I would be, there is one legitimate concern embedded in the job apocalypse narrative, and it is being used dishonestly to prop up claims far larger than it actually supports.
Young workers in AI-exposed roles are experiencing measurable employment pressure. Stanford and Yale research found a 16% relative employment decline for workers aged 22 to 25 in AI-exposed jobs since late 2022, and approximately a 20% decline specifically for young software developers. Workers aged 35 to 49 in the same sectors were up 8% over the same period.
The pattern suggests that AI is compressing the entry-level end of career ladders rather than eliminating careers wholesale.
This is a real problem and it deserves serious policy attention.
New entrants to knowledge work professions have traditionally built skills through the “grunt work” that experienced practitioners no longer need to do. If AI is automating that grunt work, the pipeline for developing the next generation of senior practitioners breaks down over a ten-to-fifteen year horizon. And that, my intelligent friend, that is a genuine structural risk worth analyzing carefully.
But it’s certainly not 40% of jobs gone and the associated white-collar bloodbath.
But what it really is, is a career ladder compression problem affecting young workers in specific high-AI-exposure sectors. It requires workforce transition investment, educational adaptation, and thoughtful policy, but what it does not require is Sam Altman telling Tucker Carlson that he is uncertain whether software engineering will exist as a profession in five to ten years, a statement which Altman made while running a company that employs thousands of software engineers and has no apparent plans to stop.
Big Tech and Big AI are the most sophisticated hypocrites money can buy, and money is currently buying a lot of them.
And Dario Amodei deserves a special mention here. This is the man who made international headlines clutching his pearls about Anthropic’s technology being used by ICE, wringing his hands about AI ethics and the dignity of vulnerable people, positioning himself as the conscience of the industry.
Yeah, noble stuff Mario. Very moving. Someone should frame it the same way you frame the AI apocalypse.
By the way, this is also the man who privately tells rooms full of CEOs that most of them are about to replace large fractions of their workforce with AI and “just don’t know it yet”, and at the same time publicly performing the augmentation gospel for journalists and regulators. That is Dario playing Harvey Dent, with two different scripts for two different audiences, and that my friends, has a shorter name in most languages which I won’t repeat in here.
The ‘threat’ Mr. Amodei is worried about is not the one he talks about in public. The actual wool-over-eyes operation is not happening at the border but is taking place in the valuation models and their earnings calls and it’s finding its way into the breathless job apocalypse predictions that conveniently require another $50 billion in infrastructure investment to navigate safely.
ICE is not borrowing $143 billion it cannot pay back while telling the public the robots are eating their jobs. Amodei’s own company is.
Bunch of hypocrites.
The real question worth asking
If you’ve made it to this paragraph, I congratulate you, my intelligent and curious friend. You are not the one that has to fear losing your job, and no, I cannot guarantee it never happens to you, but what I can give you is the knowledge that the next time a tech CEO tells you the robots are coming for your job, I would like you to hold two numbers in your head simultaneously.
The first number is 0.5%.
That is Anthropic’s revenue-to-valuation ratio. That is the company telling you that half of entry-level white-collar work will be automated within five years currently converts approximately half a cent of every dollar of its market value into actual revenue. The company telling you your job is probably temporary cannot itself demonstrate a path to profit on its current timeline using its own technology.
The second number is 7%.
That is the percentage of 2026 layoff plans that explicitly cite AI as a reason. The other 93% cite contract losses, market conditions, and restructuring, which is a euphemism for “we hired too many people when money was free and we are now correcting that while everyone is distracted by the robot panic we are encouraging”.
The fear is not wrong, but the attribution is.
Something large is happening in the technology economy for sure, but it is happening to the business models of legacy software companies, to the balance sheets of companies spending trillions on infrastructure they cannot yet monetize, and to the investment theses of funds that have bet civilization-scale capital on a technology whose productivity impact has not yet appeared in any macroeconomic dataset.
Your job is probably fine. And I say that as a person running AI automation factories at scale, and again no, I cannot guarantee everyone will keep their job when AI automates repeatable, low risk processes, but the business model of the people telling you otherwise is the thing that is genuinely under threat.
Now, I’ll end this rant with a quote.
The most expensive thing Sam Altman ever built was not a language model. It was the story about the language model. That one might actually be worth $500 billion.
Signing off,
Marco
I build AI by day and warn about it by night. I call it job security. Big Tech keeps inflating its promises, and I just bring the pins and clean up the mess.
👉 Think a friend would enjoy this too? Share the newsletter and let them join the conversation. LinkedIn, Google and the AI engines appreciates your likes by making my articles available to more readers.
To keep you doomscrolling 👇
- I may have found a solution to Vibe Coding’s technical debt problem | LinkedIn
- Shadow AI isn’t rebellion it’s office survival | LinkedIn
- Macrohard is Musk’s middle finger to Microsoft | LinkedIn
- We are in the midst of an incremental apocalypse and only the 1% are prepared | LinkedIn
- Did ChatGPT actually steal your job? (Including job risk-assessment tool) | LinkedIn
- Living in the post-human economy | LinkedIn
- Vibe Coding is gonna spawn the most braindead software generation ever | LinkedIn
- Workslop is the new office plague | LinkedIn
- The funniest comments ever left in source code | LinkedIn
- The Sloppiverse is here, and what are the consequences for writing and speaking? | LinkedIn
- OpenAI finally confesses their bots are chronic liars | LinkedIn
- Money, the final frontier. . . | LinkedIn
- Kickstarter exposed. The ultimate honeytrap for investors | LinkedIn
- China’s AI+ plan and the Manus middle finger | LinkedIn
- Autopsy of an algorithm – Is building an audience still worth it these days? | LinkedIn
- AI is screwing with your résumé and you’re letting it happen | LinkedIn
- Oops! I did it again. . . | LinkedIn
- Palantir turns your life into a spreadsheet | LinkedIn
- Another nail in the coffin – AI’s not ‘reasoning’ at all | LinkedIn
- How AI went from miracle to bubble. An interactive timeline | LinkedIn
- The day vibe coding jobs got real and half the dev world cried into their keyboards | LinkedIn
- The Buy Now – Cry Later company learns about karma | LinkedIn
Leave a Reply