AI Jobs Hysteria: What the Data Says for Sri Lankan Devs
Layoffs at Coinbase, Meta, and Cisco dominate headlines, but labour economists say the AI jobs apocalypse hasn't arrived — and one specific number should concern every fresh CS graduate.
Every second LinkedIn post right now predicts the end of knowledge work. Layoffs at Coinbase, Meta, and Cisco are being read as portents of what's coming for the rest of us. The tech-unemployment apocalypse, the story goes, is already here — just unevenly distributed.
But a reality-check piece in MIT Technology Review published today pushes back on this hard, and the findings are more interesting than either the doomscrolling or the "AI is overhyped" rebuttals. The data isn't entirely reassuring — but it is precise, and precise is useful.
📊 What the Macro Data Is Actually Saying
The big picture from current labour market data does not match the apocalypse narrative:
| Signal | Finding |
|---|---|
| Unemployment in AI-exposed occupations | Lower than in less-exposed fields |
| Companies using AI in any business function | Only 1 in 5 |
| Mass workforce shifts from "threatened" to "safer" jobs | No evidence found |
| Wages in AI-exposed sectors | Rising |
Labour economist Erika McEntarfer summarises it directly:
"Disruption is not yet here, and that we have time to plan."
The full-employment paradox is striking: the sectors most touched by AI tools are holding employment better and paying more than the rest of the market. If AI were wiping out these jobs at scale, you'd expect to see the opposite pattern — falling wages, rising unemployment, people retooling into adjacent roles. None of that is showing up in the aggregate numbers.
🔍 Where the Disruption Is Genuine
None of that means everything is fine. The article surfaces a specific, measurable trend that should get your attention if you are in your early-to-mid twenties and studying or starting a software career:
- Entry-level coding jobs for the 22–25 age group are down 16% since 2024
- Older, more experienced workers in AI-exposed roles are gaining positions
- Overall coding employment is still growing, but at a ~3% slower rate than the pre-AI baseline
- The net effect: experienced engineers absorb the AI productivity uplift; juniors absorb the hiring slowdown
So the aggregate employment number looks stable largely because experienced engineers are being retained and promoted into roles where they work alongside AI tooling. The entry-level intake is what is being squeezed.
Key takeaway: The jobs apocalypse hasn't arrived, but the apprenticeship pipeline is narrowing. Fewer junior positions means the entry gate for a software career just got tighter.
This makes sense if you think about how teams actually use tools like GitHub Copilot. A senior developer with an AI coding assistant can now handle a sprint's worth of boilerplate that would previously require a junior ramping over three to six months. The junior role isn't gone — there are just fewer of them per senior.
🕰️ Why Historical Patterns Cut Both Ways
The Technology Review piece flags a useful track record: past predictions about AI-driven job obliteration consistently missed.
- Radiologists were supposed to be replaced by deep-learning diagnostics within a few years of AlexNet's 2012 breakthrough. Radiology employment has since grown.
- Truck drivers have faced full-automation predictions for a decade. Long-haul autonomous freight still hasn't arrived at scale.
The pattern is that technology augments and reshapes roles more than it eliminates them outright. The problem is that "reshaping" still causes real hardship for individuals who can't adapt quickly, even when aggregate employment looks healthy.
A job redefinition that leaves total headcount unchanged can still gut the opportunities for anyone entering the field — because the on-ramp, not the destination, is where the narrowing happens.
This is a distribution problem more than a total-employment problem. It is precisely why the age-tier split — experienced workers gaining, juniors losing — matters more than any single headline about layoffs at a named company.
💡 What This Means for Sri Lankan CS Students and Engineers
The 16% drop in entry-level hiring is not a statistic you can safely ignore because you are in Colombo rather than San Francisco. The companies hiring Sri Lankan developers remotely — and the local firms that benchmark against global tech hiring — are tracking the same market. A few things follow:
1. The portfolio bar just moved up.
Employers who previously hired juniors for boilerplate work now have AI tools for that. What they need from a new hire is evidence that you can define a problem and ship a solution end-to-end, not just execute instructions. Your GitHub profile needs to show something you built and deployed — a working tool, a live API, a dataset pipeline — not just course completion certificates.
2. AI tool fluency is now table stakes.
If you are still planning to "learn AI" as a future goal, you are already behind the minimum. The actually useful skill is knowing when to trust the output, when to debug it, and when to throw it away entirely. That is applied judgment, and it is what teams are paying for.
3. Experienced developers have more to gain than to fear.
The data shows wages in AI-exposed sectors are rising. If you have several real codebases behind you, you are the human judgment layer that makes AI tooling productive. The worst likely outcome for you is a slower-than-expected salary jump, not redundancy. If you are in this position, it is worth modelling what a salary bump actually means for your take-home after APIT — our [Sri Lanka [Income Tax](https://induwara.lk/tools/tax-calculator-global) Calculator](https://induwara.lk/tools/sri-lanka-tax-calculator) runs the current 2025/26 IRD brackets instantly.
4. The local market has a different threat model.
Software development for Sri Lankan banking, government, and SME markets requires local domain knowledge and relationships that remote AI automation cannot easily substitute. If you are building tools for local businesses — even on a freelance basis — the competitive pressure from AI is much lower than the global headlines imply.
What This Means for You
The honest summary: the AI jobs apocalypse is not here. Employed knowledge workers are mostly safe, and wages for experienced developers are improving. But one thing is clearly shifting in a way that matters: the entry-level pipeline is narrowing, and the 22–25 cohort is absorbing the early pain.
If you are a final-year student, a recent graduate, or someone building a small tech team in Sri Lanka, the implication is the same — treat AI fluency as your baseline capability, not your differentiator. The developers who will feel this most sharply are the ones waiting for things to settle before adapting.
And for everyone reading the layoff headlines: stop treating each named company's restructuring as aggregate proof of collapse. Look at the actual labour data. The story is more boring, more specific, and ultimately more actionable than the apocalypse narrative — and "more actionable" is the more useful thing to be working with right now.
Original source
A reality check on the AI jobs hysteria