(Long article but worth reading) Your Job Won’t Be Automated. It Will Be Deleted.

[Flibbertigibbet]

Your Job Won’t Be Automated. It Will Be Deleted.​

Paweł Targosiński
16 min read
2 days ago

Press enter or click to view image in full size

Source: Image generated by the author using ChatGPT

AI isn’t transforming the labor market. It’s dismantling it. Industry by industry. Job by job. And nobody is warning you in time.

I’ve been running a software company for 15 years. I build technology for brands like DHL, Nestlé, Sii and Samsung. Every day, I see firsthand what new AI tools can do. And I need to be blunt: what’s coming will reshape the lives of millions of people. Most of them will find out too late.

I’m not writing this as a futurist. I’m writing this as a practitioner​

I’m not a researcher at MIT or an analyst on Wall Street. I’m a tech entrepreneur from Poland who works with these tools daily. That’s exactly why I’m writing this, because I see a dangerous gap between what’s happening inside tech companies and what reaches the general public.

When ChatGPT launched in 2022, most people treated it as a novelty. A toy that writes poems and answers trivia. Three years later, that “toy” writes code at a mid-level engineer’s ability, generates video from text descriptions, diagnoses diseases from X-ray images, and handles customer conversations in 35 languages simultaneously.

And this is just the beginning.

The numbers nobody is showing you on the evening news​

In 2025, U.S. companies laid off nearly 55,000 people explicitly citing AI as the reason. That’s twelve times more than just two years earlier [1]. Total U.S. layoffs in 2025 hit 1.17 million, the highest since the COVID-19 pandemic [2].

But the real story isn’t in aggregate numbers. It’s in the specific companies, and how casually they’re doing it.

Amazon cut 14,000 corporate jobs in October 2025, then added another 16,000 in early 2026. CEO Andy Jassy wrote in an internal memo that the company expects fewer people in roles that exist today and more AI agents handling the work [3].

Microsoft eliminated 15,000 jobs through 2025. CEO Satya Nadella revealed that 30% of the company’s code is now written by AI. Over 40% of the layoffs targeted software engineers [4].

Salesforce slashed its customer support division from 9,000 to 5,000 people using AI agents. CEO Marc Benioff confirmed this publicly [5].

Klarna, the European fintech giant, shrank from 5,000 to 3,000 employees. Their AI chatbot handled two-thirds of all customer conversations, equivalent to the work of 853 full-time agents, saving the company $60 million [6]. When quality dropped and customers complained, Klarna partially reversed course, rehiring some humans. But they never went back to the old headcount. The new model is a hybrid: AI handles simple queries, humans only take the complex ones. New hires work in a flexible, “Uber-style” setup with lower pay. This is the template for what’s coming across every industry: not instant replacement, but gradual compression. Fewer people. Lower wages. AI doing most of the routine work.

IBM deployed its AskHR system, which handles 11.5 million interactions annually with minimal human oversight, replacing hundreds of HR workers [7].

HP announced 4,000 to 6,000 job cuts as part of an AI productivity initiative targeting $1 billion in savings [8]. Chegg, an education platform, fired 45% of its workforce citing “the new realities of AI” [9]. Duolingo declared itself “AI-first” and immediately cut 10% of its contractors, with AI taking over translations across 100+ languages [10]. UPS eliminated 20,000 positions, one of the largest reductions in its 116-year history [11]. Workday cut 1,750 jobs (8.5% of staff) to redirect resources toward AI investment [12].

These aren’t predictions. This has already happened.

A September 2025 survey by Resume.org found that 37% of companies expect to have replaced jobs with AI by end of 2026 [13]. Over half of business leaders surveyed admitted they had already conducted layoffs in 2025. And according to Axios, every single CEO they spoke with is working furiously to figure out when and how AI agents can displace workers at scale. The public, they noted, will only realize it when it’s too late [14].

Every new AI product announcement is becoming an existential threat to entire industries​

Investors understand something that most workers don’t yet see: every new AI tool released is a potential replacement for thousands of jobs overnight.

On February 20, 2026, Anthropic, the company behind the Claude AI, announced a new tool called Claude Code Security. It scans codebases and autonomously finds security vulnerabilities, doing the work previously handled by specialized cybersecurity firms. Within hours, major cybersecurity companies lost billions in combined value. One product, from one company, targeting one niche [15].

This pattern keeps repeating. Earlier in February 2026, Anthropic launched another product (Claude Cowork plugins) and the enterprise software sector experienced one of its worst selloffs in years, approaching declines not seen since 2008 [16].

Mustafa Suleyman, head of Microsoft AI, predicted that most tasks performed by accountants, lawyers, and other white-collar specialists will be fully automated within 12 to 18 months [17]. Even if he’s exaggerating and it takes 18 to 36 months, the scale of the change will be enormous. And as one tech commentator put it: “This cart is moving too fast. Nobody can stop it. And nobody has any idea how much the crash into the wall is going to hurt.”

When an AI startup can threaten the business model of an entire industry with a single product launch, the message is clear: no sector is safe.

Industry by industry: who’s on the chopping block​

This isn’t about some abstract future. These are specific industries where changes are happening right now or will hit within 2–5 years.

Software development​

I’ll start with my own industry, because the changes here are most visible.

GitHub Copilot, Cursor, Claude Code: these tools don’t yet replace senior engineers, but they drastically reduce the need for juniors. Meta announced that their AI can effectively replace a mid-level engineer. Shortly after, they laid off 5% of their workforce [18].

There’s another angle that people outside the industry don’t see: AI now writes tests. Unit tests, integration tests, code reviews. This was traditionally the bread and butter of junior developers, the first real task they’d get on a team. AI can generate them in seconds. It can even write tests designed to pass, which, ironically, is also what juniors sometimes do. Except AI does it faster and cheaper.

In practice, a team that needed 10 developers now needs 4, with AI acting as the fifth, sixth, and seventh.

Customer service and call centers​

This is the industry AI hit fastest and hardest. Sam Altman, CEO of OpenAI, stated he’s “confident” AI will replace a significant number of customer support roles [19]. The World Economic Forum flagged customer service agents, administrative assistants, and bank tellers as among the most vulnerable occupations [20].

Accounting and finance​

AI can already analyze invoices, file taxes, generate financial reports, identify anomalies, and conduct audits. Major auditing firms are testing the replacement of juniors with algorithms. JP Morgan Chase openly discusses AI as a factor reducing back-office employment [21].

The point isn’t that AI will replace every accountant. The point is that an office that employed 20 people will need 5, because the other 15 were doing work that algorithms handle faster and more accurately.

Graphic design​

Midjourney, DALL-E, Adobe Firefly, Stable Diffusion. These tools generate logos, interior designs, product visualizations in seconds. Freelancers on Fiverr and Upwork are already feeling it. Why pay a designer $500 for a project when AI generates 50 variants in a minute? Not all will be perfect, but for 80% of business needs, they’re good enough.

Ad agencies that once employed teams of designers to create ad variants now use AI to generate hundreds of versions in hours.

Film and video​

The 2023 Hollywood actors’ and writers’ strike wasn’t a coincidence. It was the first alarm bell.

Tools like Sora (OpenAI), Runway, and Kling generate video content that would have looked like science fiction a year ago. McKinsey estimates that by 2030, AI could handle $10 billion worth of U.S. content production [22]. Morgan Stanley warns that generative AI is available to both major studios and new entrants, meaning that within five years, individuals will be able to produce professional-quality output [23].

A feature film without actors? A music video without a camera? This is a matter of months, not years.

Music​

Platforms like Suno and Udio generate complete songs, with vocals, instruments, and full arrangements, from a few words of description. Voice cloning technology can create convincing replicas of any singer’s voice. The viral track “Heart on My Sleeve,” which used deepfaked voices of Drake and The Weeknd, demonstrated this to the world [24].

Copywriting and content​

This battle is already lost. Most simple website copy, product descriptions, and social media posts: AI writes them faster, cheaper, and often better than average copywriters. Studies show over 80% of digital marketers fear AI will replace content writers [25]. That fear is becoming reality.

Translation​

DeepL and AI-integrated language models translate at levels that were unreachable three years ago. Duolingo is the symbolic example: a company that taught people languages fired its translators because AI does it better. Translation agencies worldwide are reducing teams.

Law​

AI analyzes contracts, searches for precedents, prepares briefs, and verifies document compliance. Law firms are discovering they can replace entire research teams with software subscriptions. Young lawyers who just finished their training are entering a market where their core work is being automated [26].

Medicine​

AI already outperforms doctors in many imaging diagnostics: X-rays, CT scans, dermatology, pathology [27]. Medical speech recognition systems transcribe doctor-patient conversations with near-perfect accuracy.

It won’t replace surgeons. But radiologists, pathologists, dermatologists? That’s a real threat within 3–5 years. Geoffrey Hinton believes healthcare could benefit overall because AI will expand access to diagnostics. But that means fewer doctors needed for the same volume of work [28].

Transportation and logistics​

Self-driving cars (Waymo, Tesla), delivery drones, automated warehouses. Amazon already operates warehouses staffed primarily by robots. Salman Khan of Khan Academy predicts autonomous vehicles will become the norm, potentially impacting over one million drivers in the U.S. alone [29].

Manufacturing and humanoid robots​

This is the element that changes everything.

“Dark factories,” facilities that operate without lights or people, are already running in China and Japan. But what’s coming is far more powerful.

In early 2026, the humanoid robotics industry hit an inflection point. Tesla began mass-producing Optimus Gen 3 at its Fremont factory, the same facility where it previously made Model S and Model X vehicles. Those car lines were shut down to make room for robots. Musk is targeting a price below $20,000 per unit and production capacity of one million units per year [30].

Tesla isn’t alone. Boston Dynamics shipped Atlas robots to Hyundai’s Georgia factory for real production work. Figure AI launched its BotQ facility capable of producing 12,000 Figure 03 robots annually. Agility Robotics deployed Digit robots at Toyota Canada. 1X Technologies started delivering NEO home robots at $20,000. Unitree sells a humanoid robot for $5,900 [31].

Projections indicate tens of thousands of humanoid robots deployed globally by the end of 2026, and over 100,000 cumulatively by 2027. China controls 85–90% of the market with over 140 companies in the space [32].

The combination of humanoid robots with AI creates something unprecedented: physical automation that learns, adapts, and improves every day. A robot that doesn’t get sick, doesn’t take vacation, works 24 hours a day, 365 days a year. For a one-time cost equal to a few months of an employee’s salary.

Why this time is different​

I hear the argument all the time: “It’s always been this way. Machines took jobs from weavers but gave them factories. Computers took jobs from typists but gave them offices. It’ll be the same.”

No. It won’t. Here’s why.

Speed. The Industrial Revolution took decades. Office computerization took a generation. AI is entering dozens of industries simultaneously, and the pace accelerates every quarter. The gap between GPT-3 and today’s models spans just a few years, yet the difference in capability is like that between a calculator and a computer.

Breadth of impact. Previous technological revolutions hit one category of work: physical or simple cognitive. AI strikes mental, creative, and physical work all at once. There is no obvious “next place” for people to move to.

The middle class is the target. This is the most dangerous aspect. Previous automation hit the working class. AI hits white-collar workers: people with degrees, mortgages, a sense of security. Lawyers, accountants, designers, programmers, radiologists. People who believed education would protect them.

No safety net. Education systems, governments, and social institutions operate in decade-long cycles. AI changes in quarterly cycles. Most countries have no serious public discussion about what’s coming.

What the people who build this technology are saying​

You don’t have to take my word for it. Listen to the people who create this technology.

Dario Amodei, CEO of Anthropic: “We, as the producers of this technology, have a duty and an obligation to be honest about what is coming. I don’t think this is on people’s radar.” His scenario: cancer cured, economy growing at 10% per year, budget balanced, and 20% of people without jobs [33].

Geoffrey Hinton, Nobel laureate, “godfather of AI”: “It seems very likely to a large number of people that we will get massive unemployment caused by AI.” He acknowledged new jobs will emerge but doesn’t expect the number to come close to those eliminated. He dismissed universal basic income as inadequate because it doesn’t address the loss of dignity and purpose tied to work [28].

Bill Gates: humans may soon not be needed “for most things” [34].

Elon Musk: most humans won’t have to work at all “in less than 20 years” [35].

Salman Khan, CEO of Khan Academy: the AI revolution will hit “faster and harder than anyone is anticipating.” Even a 10% reduction in employment “will feel like a depression” [29].

Get Paweł Targosiński’s stories in your inbox​

Join Medium for free to get updates from this writer.
World Economic Forum: 41% of employers plan to reduce their workforce by 2030. The IMF warns that 60% of jobs in advanced economies are already exposed to AI [20]. The Inter-American Development Bank estimates 980 million jobs worldwide face high disruption risk within the next year [36].

Senator Bernie Sanders estimated that nearly 100 million U.S. jobs could be displaced by automation. He wrote: “It’s not just economics. Work, whether being a janitor or a brain surgeon, is an integral part of being human. What happens when that vital aspect of human existence is removed from our lives?” [37].

The coming inequality​

The problem isn’t just job loss. It’s what happens to society when the “safe” middle class starts losing ground.

The World Economic Forum warns of what they call the “AI precariat”: a new class of people who will lose not only their jobs but their identity and sense of purpose [20]. The homeless people you see in New York City could become a global phenomenon. With one difference: the people on the street will be those who recently had homes, cars, and retirement plans.

The Forrester research group found that 55% of companies regret laying off workers for AI that wasn’t ready yet. But instead of admitting the mistake and rehiring at previous salaries, they quietly fill gaps with cheaper offshore labor [38]. Meanwhile, only 16% of workers had high AI readiness in 2025. Gen Z has the highest readiness at 22%, yet companies are eliminating the entry-level positions that would bring them into the workforce [39].

Some argue this is all overblown. Oxford Economics suggests that AI-related layoffs are sometimes “dressing up standard cuts as a positive story for investors” [40]. The Economist writes that unemployment is low and wages are still growing. They may be right. Today. But disruptive technologies don’t move linearly. Nothing changes for a long time, and then everything changes at once.

What you can do. A concrete plan.​

I’m not writing this to paralyze you with fear. I’m writing this so you have time to act. Awareness of what’s coming is your biggest advantage, but only if you turn it into concrete steps.

1. Learn to use AI. Today.​

Not next year. Not “someday.” Today. Install ChatGPT, Claude, Gemini. Start using them in your daily work. You don’t need to code. You need to learn how to ask good questions and verify answers. People who master this will be 3–5x more productive than those who don’t. Data already shows that graduates with greater AI skill exposure earn more and find jobs faster [39].

2. Diversify your income.​

Don’t depend on a single source. If your only income is a salary, start thinking about additional streams. Rental property. A small online business. Services that are hard to automate. The more legs your financial stool has, the harder it is to knock over.

3. Build a cash reserve.​

Minimum 6–12 months of living expenses in liquid savings. In a crisis, cash buys you time to think.

4. Avoid taking on major new debt.​

A 30-year mortgage that assumes your industry will exist in its current form? Think twice. This doesn’t mean “don’t buy a home.” It means “don’t assume your income will be stable for decades.”

5. Invest in skills AI can’t easily replace.​

Human relationships. Negotiation. People management. Creative problem-solving for complex, multi-dimensional challenges. Empathy. Leadership. And paradoxically: trades that require physical presence. Plumbers, electricians, mechanics, nurses. These jobs may prove safer than “prestigious” office positions.

6. Talk to your kids about this.​

Don’t tell them “study law, you’ll have a good career.” Don’t tell them “study computer science, you’ll be set.” Teach them systems thinking, adaptability, and how to work with technology. A specific profession may not exist in 10 years. The ability to learn will.

7. Build real relationships.​

Not on LinkedIn. Not on Facebook. In real life. A network of people you trust with complementary skills. In every crisis, from wars to recessions, the people who fared best were those with strong social support. That’s insurance you can’t buy.

8. Be on the side of people deploying AI.​

This sounds cynical, but it’s realistic advice. In every technological revolution, the winners were those who understood it and leaned in, not those who fought the inevitable. You don’t need to build AI. You need to understand how to leverage it in your field.

If you run a business or are thinking about starting one, consider where the opportunities are. During a gold rush, the people who got rich weren’t the miners. They were the ones selling shovels. The same logic applies here.

Companies across every sector are desperate to figure out AI. Most don’t know where to start. If you can become the person or company that helps them implement it, train their teams, or redesign their workflows around AI, you’re positioned on the winning side of this shift. AI consulting, implementation services, training programs, building custom AI-powered tools for specific industries: these are the “shovels” of the AI gold rush.

Think about niches that are underserved. A local accounting firm doesn’t need OpenAI. They need someone who understands both accounting and AI well enough to set up automated invoice processing, tax analysis, and anomaly detection for their specific workflows. A construction company doesn’t need a PhD in machine learning. They need practical help integrating AI into project estimation, safety monitoring, or supply chain management.

Another angle: create products or services that serve people displaced by AI. Reskilling platforms, career transition coaching, mental health support for professionals losing their identity along with their job. The World Economic Forum calls this the “AI precariat” problem, and almost nobody is building solutions for it yet.

The key principle is simple: position yourself where the wave is going, not where it’s been. Every industry will need people who bridge the gap between raw AI capability and real-world application. That bridge is where the value will concentrate.

I don’t want to be a prophet of doom​

Maybe I’m wrong about the timing. Maybe governments will respond faster than I expect. Maybe entirely new industries will emerge that we can’t yet imagine.

But I’d rather warn you too early than too late.

Because one thing I know for certain: this change is coming. The only questions are how fast and who will be ready.

As Geoffrey Hinton put it: “It’s a bit like when you drive in fog. You can see clearly for 100 yards and at 200 yards you can see nothing. We can see clearly for a year or two, but 10 years out, we have no idea what’s going to happen” [28].

I’m not telling you this so you live in fear. I’m telling you so you have time to prepare.

Because the only thing worse than a crisis is a crisis that catches you off guard.
If this was valuable to you, share it. Not for reach. For the people you know who might not see what’s coming.

References​

[1] Challenger, Gray & Christmas, “AI-Related Job Cuts Report 2025,” 2025. [Online]. Available: https://www.challengergray.com/

[2] Bureau of Labor Statistics, “Mass Layoff Statistics 2025,” U.S. Department of Labor, 2025.

[3] E. Kim, “Amazon to cut 14,000 corporate roles amid AI push,” Business Insider, Oct. 2025.

[4] S. Nadella, Microsoft Q4 2025 Earnings Call transcript, Microsoft Corporation, 2025.

[5] M. Benioff, “Salesforce AI Agents Replace Customer Support Roles,” Salesforce Earnings Call, 2025.

[6] Klarna, “Klarna AI Assistant Handles Two-Thirds of Customer Service Chats,” Klarna Press Release, 2024. [Online]. Available: https://www.klarna.com/

[7] IBM, “IBM AskHR: Transforming Human Resources with AI,” IBM Case Study, 2025. [Online]. Available: https://www.ibm.com/

[8] HP Inc., “HP Future Ready Transformation Plan,” HP Press Release, 2025.

[9] Chegg Inc., “Chegg Announces Workforce Restructuring,” Chegg Press Release, 2025.

[10] Duolingo, “Duolingo Shifts to AI-First Strategy,” Duolingo Blog, 2025. [Online]. Available: https://blog.duolingo.com/

[11] UPS, “UPS Announces Organizational Changes,” UPS Press Release, 2025.

[12] Workday, “Workday Announces Restructuring to Accelerate AI Investment,” Workday Press Release, 2025.

[13] Resume.org, “AI Replacing Jobs Survey: 37% of Companies Expect Replacement by 2026,” Resume.org, Sep. 2025.

[14] Axios, “CEOs Quietly Planning AI Worker Displacement,” Axios, 2025. [Online]. Available: https://www.axios.com/

[15] Anthropic, “Introducing Claude Code Security,” Anthropic Blog, Feb. 20, 2026. [Online]. Available: https://www.anthropic.com/

[16] Reuters, “Enterprise Software Stocks Slide on AI Competition Fears,” Reuters, Feb. 2026.

[17] M. Suleyman, Interview on AI and White-Collar Automation, Microsoft AI, 2025.

[18] Meta Platforms, “Meta Announces Workforce Reduction,” Meta Newsroom, 2025.

[19] S. Altman, Interview on AI and Customer Service, OpenAI, 2025.

[20] World Economic Forum, “Future of Jobs Report 2025,” WEF, Jan. 2025. [Online]. Available: https://www.weforum.org/reports/the-future-of-jobs-report-2025/

[21] JP Morgan Chase, “Annual Report 2025: AI and Operational Efficiency,” JPMorgan Chase & Co., 2025.

[22] McKinsey & Company, “The Economic Potential of Generative AI,” McKinsey Global Institute, 2025.

[23] Morgan Stanley, “Generative AI and the Future of Content Production,” Morgan Stanley Research, 2025.

[24] B. Sisario, “An A.I. Hit: How a Fake Drake Song Exposed the Music Industry,” The New York Times, Apr. 2023.

[25] Semrush, “State of Content Marketing 2025: AI Impact Survey,” Semrush, 2025.

[26] Thomson Reuters, “AI and the Legal Profession: 2025 Report,” Thomson Reuters Institute, 2025.

[27] The Lancet Digital Health, “AI Performance in Medical Imaging: A Systematic Review,” vol. 7, 2025.

[28] G. Hinton, Interview on AI risks and unemployment, Nobel Prize Press Conference, 2024.

[29] S. Khan, “AI, Education and the Future of Work,” Khan Academy Blog, 2025.

[30] Tesla, “Optimus Gen 3 Production Update,” Tesla Investor Relations, Jan. 2026.

[31] IEEE Spectrum, “The Humanoid Robot Race: 2026 Industry Overview,” IEEE Spectrum, Feb. 2026.

[32] International Federation of Robotics, “World Robotics Report 2025,” IFR, 2025.

[33] D. Amodei, “Machines of Loving Grace,” Anthropic Blog, Oct. 2024. [Online]. Available: https://www.anthropic.com/

[34] B. Gates, “AI Is About to Completely Change How You Use Computers,” GatesNotes, 2025. [Online]. Available: https://www.gatesnotes.com/

[35] E. Musk, Interview at VivaTech Paris, May 2024.

[36] Inter-American Development Bank, “AI and Labor Markets in Latin America,” IDB Report, 2025.

[37] B. Sanders, “We Must Not Let the Billionaire Class Use AI to Destroy the Working Class,” The Guardian, 2025.

[38] Forrester Research, “The State of AI-Driven Workforce Decisions,” Forrester, 2025.

[39] Forrester Research, “AI Readiness Index: Worker Preparedness Report 2025,” Forrester, 2025.

[40] Oxford Economics, “AI and Employment: Separating Hype from Reality,” Oxford Economics Research Brief, 2025.

 

[Flibbertigibbet]

I am wondering...

So when the senior workers retired in next 5 years, then no or few new entry workers because the jobs replaced by AI, the employment structure of a company will transformed from pyramid shape to diamond shape?

 

[True Believer]

The management and staff at my company from age 55 and older seem to be getting phased out every year. Even if they are not laid off, they are unfairly appraised as average to poor performers to save on promotion, increments and bonuses. The company knows older staff have few options except to stay on if they need the job. Those who have saved enough can voluntarily take early retirement.