Day 993: The Jobpocalypse Has Begun
Something is happening to your job. The data says so.
Free post — The Next 1000 Days
Something is happening to your job.
You might not feel it yet.
But the numbers don’t care what you feel.
The names you recognise are already moving
Last year, US employers announced over 1.17 million job cuts — the highest total since the Covid-19 pandemic in 2020 (CNBC). Of those, about 55,000 were explicitly linked to AI adoption, according to Challenger, Gray & Christmas (Equitypandit). That sounds manageable, right? A rounding error in a 160-million-person workforce.
Here is the problem with that framing: CFO survey data suggests AI-related job cuts could be nine times higher in 2026 than in 2025 (Fortune).
Nine times.
The dam isn’t broken. But the cracks are spreading fast, and almost no one in charge is watching.
This isn’t speculative. The announcements are public. Amazon announced the largest round of layoffs in its history, slashing 14,000 corporate roles and explicitly citing AI as the driver. Salesforce CEO Marc Benioff confirmed he cut customer support from 9,000 workers to 5,000 — because AI was already doing up to 50% of the work (CNBC). Meanwhile Microsoft cut 15,000 jobs, and Oracle cut thousands more — both framing the layoffs as their businesses “doing more with less” through AI (TechRepublic). Microsoft reported quarterly revenue of $70.1 billion, up 13% year-on-year (TechRepublic). They are not struggling. They are thriving. They just need fewer of us to do it.
IBM’s CEO confirmed that AI chatbots took over the jobs of several hundred HR workers (CNBC). Not the jobs of the future — the jobs that existed last Tuesday.
Who gets hit first
If you are early in your career, the data is brutal. Job listings for entry-level corporate roles have declined 15% over the past year, and over the past two years, there has been a 400% increase in employers using “AI” in job descriptions, according to career platform Handshake (CBS News). Translation: they want one person who can direct AI, not five people to do the work.
The brunt of job losses is falling on entry-level roles — data entry, customer service, admin, help desk. Businesses are not just laying people off; they are eliminating the roles entirely (TechRepublic).
And it’s not only the obvious targets. In early 2026 alone, there were 32,000 job losses in technology firms. Nearly 55,000 job cuts were directly attributed to AI across 2025 — and over 75% of those happened after 2023 (AIMultiple).
The curve is accelerating, not flattening.
Here is what they’re not telling you
Every AI company will tell you they want to augment workers, not replace them. They will use words like “partnership” and “co-pilot.” That’s the $20-a-month story — the one where you buy a subscription and become superhuman.
The real story is different. The actual product these companies are building is your replacement, sold to your employer for a few thousand dollars a month. Not a tool that helps you — a service that removes the need for you. The incentives are not aligned with your job security. They never were.
There is no government committee managing this transition. No international body setting a humane pace. No adult in the room deciding that 3% annual job displacement is the safe speed, not 30%. The decision theorist Eliezer Yudkowsky put it bluntly: in a sane world, we would coordinate to ensure the pace of automation doesn’t outrun our ability to adapt. We don’t live in that world. We live in this one, where capability races ahead and the workforce scrambles to catch up.
Some economists will tell you new jobs will emerge. Historically, they’re right — the steam engine and the internet both created more jobs than they destroyed. But that transition took decades, and it happened when the technology had hard limits. Today’s models have no obvious ceiling. Anthropic’s new Claude Mythos model has already identified thousands of software vulnerabilities across every major operating system and browser — a capability previously thought to require elite, state-sponsored hacking teams (The Atlantic/DNYUZ). Anthropic considered it too powerful for public release (AOL/Anthropic). That’s the version they’re keeping back. Think about what it says about the versions they’re shipping.
On software engineering benchmarks, Mythos jumped from 80.8% to 93.9% — and on high-difficulty mathematical reasoning, from 42.3% to 97.6% (36kr). That is not incremental progress. That is a step change, in a single model generation.
Speaking personally: I use Claude Code daily in my development work. It is extraordinary. Some days I feel less like a software developer and more like a code reviewer — checking what the AI built, nudging it, redirecting it. My output is up. My team size requirement is down. I am not complaining. But I am paying attention.
What about UBI?
Universal Basic Income gets floated as the answer every time this conversation comes up. Maybe it happens, maybe it doesn’t — the political obstacles are enormous, and the timeline is anyone’s guess. But here’s the thing even the optimists miss: UBI isn’t a solution, it’s a consolation prize. The problem with depending on UBI isn’t the money. It’s the leverage.
Leverage comes from contributing something the world needs. When your income is a government transfer rather than a wage, you have no seat at the table — not in your company, not in your industry, not in the economy. You become a beneficiary, not a participant. That is not a position of strength. That is the definition of dependency.
The only durable answer is to build assets that earn whether or not you are employed. Investments that compound. Income streams that don’t require your employer’s approval to exist. Financial independence is not a retirement goal for the comfortably-off. It is quickly becoming the only genuine hedge most knowledge workers have.
What to watch
I made a call in the last post on unemployment figures by year-end. Here’s what to track over the coming months:
The official unemployment rate is a lagging indicator — it will look fine right until it doesn’t. What to watch instead: entry-level job postings (are they continuing to fall?), the proportion of layoff announcements that cite AI (it was 4.5% of all cuts in 2025 — by January 2026 it had already risen to 7%, per HIGH5 Test), and whether we start to see college graduates unable to find starting roles at scale. That last one will be the canary.
The transition won’t announce itself. It will arrive as a series of ordinary-looking events — a hiring freeze here, a restructure there — until one day the aggregate is undeniable.
Don’t wait for undeniable.
The Next 1000 Days tracks one person’s attempt to build financial resilience before the labour market changes in ways we can’t fully predict. If this resonated, consider sharing it with someone who needs to read it.
Nothing here is financial advice. Any capital you lose acting on my reasoning is your problem. Any gains are obviously due to my outstanding insights.