I Lost My Job Because of AI

It landed in my calendar like a raven tapping at the window.

A 15 minute time slot scheduled for later that day with my manager whom I rarely speak with. There was no meeting description. No Slack message to say “hey just wanna catch up on…”. Just a casual meeting invite sitting there like a vulture perching on a headstone. I knew my time was up.

I rarely spoke with this manager because I’m frequently switching teams and placed under other engineering managers within the company. This isn’t all that unusual for contractors, so I figured if he wanted a random 15-minute chat out of nowhere with no indication of the topic, it could only mean bad news.

I received a message later in the day from another colleague:

“Tons of contractors getting let go today”

There was a silent culling going on, happening behind closed doors and in private slack channels. If you looked around everything else seemed normal. That morning we were knee deep in a planning session for a new integration architecture for the company. Throughout the meeting I kept bringing up various long term trade-offs and short term headwinds with our strategy, which in hindsight was very ironic. “Hhhmm yes, i think in the short term this might be a problem but if we look a year or two out then we’ll be in a better position.”

Another colleague messaged me to deliver some news

“John has just been laid off”

Minutes passed

“Mark has been laid off!”

Another few minutes passed and he wrote:

“I’VE just been laid off”

One by one, developers were being picked off like players in Squid Game.

As a contractor, I’m always expecting to be one of the first to go. Since 2022 I’ve been on high alert as it was becoming obvious where the industry was going. I’ve lived every day for the past 3 years in the knowledge that any moment my life could change in an instant. I travel a lot, so almost every Airbnb I reserved in advance had to be refundable. I hoarded money like a squirrel hoarding nuts for the winter, and managed to trick myself into thinking I wasn’t being paid as much as I was.

My parents would tell me “Oh go on…treat yourself to that thing, you can easily afford it on your salary”

“No I can’t afford it” I’d reply.

Mentally and financially preparing for this eventuality softens the blow to some extent, but it’s still a blow. My girlfriend and I had planned trips together later this year, that now look to be in doubt. The company had only just recently extended my contract, extending my tenure until March 2026. With that I felt a sense of certainty that allowed me to plan at least 8 months ahead. The company had also performed another large round of layoffs just back in January. With the layoffs earlier this year and the recent extension of my contract, this felt like a kick in the teeth and a sign that the overall financial picture of the company was bleaker than I thought.

Since 2023 layoffs had become an annual tradition within the company, with more in 2024 and again in 2025. The difference this year was the language and just how normalised this had become.

In 2023 the linguistic differences were stark.

“Team we know this is difficult for everybody. Take the time you need and i’m here if you need to talk”

“Reach out to your colleagues, take some time to process this. We’re here for you”

Fast forward to 2025 and the announcement from one of the company leaders seemed devoid of any human touch whatsoever.

“We’re in a period of economic uncertainty. Our company’s top focus remains squarely on revenue growth and improving EBITDA”

It continued

“We’ve decided to make targeted contractor reductions — an immediate lever to decrease our spending.”

20 developers let go in the space of a day, some of whom had been at the company for years, and the announcement resembled the language of a dispassionate mechanic describing which parts to remove from an engine; sanitised bloodless corporate speak designed to obscure the human impact behind executive decision making. In the space of 2 years, we’ve gone from being “humans” to mechanical parts to be disposed of.

Now in all fairness the CEO was kind enough to reach out personally to some of us who had been around for a few years, and not just with some bland generic corporate spiel. A simple Slack message showing a little remorse and some basic humanity can make a meaningful difference in how people feel during a difficult time. I also appreciated how my manager got on a call to deliver the news, as I know that’s not an easy thing to do and it’s certainly better than how some companies handle it, like cutting off access without warning and deactivating your badge.

AI Is Coming For Your Job

Artificial intelligence has been lurking in the background of our digital lives for decades. It handles many tasks like quietly filtering spam, routing traffic, and making sure your GPS doesn’t send you through a lake. But nobody wrote dystopian editorials about it until Generative AI arrived onto the scene via ChatGPT. Whilst this technology is incredibly impressive, when you distill the essence of a Large Language Model into its rawest form — next token prediction — the whole narrative becomes significantly less apocalyptic.

In the rest of this article I’ll be using the terms “AI” and “next-token prediction” interchangeably to emphasise the point.

The media loves a good AI scare story. Fear sells and generates clicks. They simply can’t help themselves. Where there’s a simple narrative that conveniently aligns with corporate interests then you know the media will rush to it like cats to milk. Never mind that the ‘AI’ in question might just be a particularly ambitious Excel macro, or that the company was already planning layoffs since Q2, or the AI was Actually Indians all along.

Who could forget the story below, which was reported all over the place

It’s like journalistic comfort food — predictable, easy to digest, and requires absolutely no critical thinking.

They frame these stories as harbingers of robotic domination instead of recognising that Andy from HR got eliminated because executives discovered they could rebrand routine downsizing as “AI-driven workforce transformation”, similarly to how “return to office” mandates were often just layoffs with extra steps, designed to shed employees without paying severance.

It’s the perfect storm of technological anxiety and corporate bootlicking where the narrative practically gift-wraps itself: humans are obsolete, resistance is futile, and isn’t it wonderful how efficiently capital can now dispose of labour? It’s disaster porn for the digital age.

So was I laid off because of AI?

Yes and no.

No I was not laid off because a next token predictor replaced me.

When we think about employees being laid off because of AI we imagine a new business process put in place, powered by auspicious new technology, that manages to complete the many tasks that form a job much more efficiently than human staff.

In my profession of software development, we can clearly see how well LLMs perform at writing code. The speed at which they can deliver is remarkable and I can now generate bugs much faster than before. Coding has become much less mundane and arduous than it was a few years back. I was already at the point in my career where I just didn’t want to manually write out another 15 unit tests again, or another HTTP client with the necessary logging and error handling.

Coding can be fun, but it can also be incredibly tedious and repetitive. Software developers have been trying to automate the production of software for decades. Makefiles, linters, code scaffolding, library dependency file generators, Swagger docs and various other tools have demonstrated just how enthusiastic software developers are of automating the creation of software.

But my job isn’t to sit down for 8 hours a day and churn out code like I’m working in some kind of digital widget factory. A software developers job is not as a flesh based IDE that cosplays as a human compiler. In fact in the lead up to my contract termination I’d hardly written any code for the entire month. I was knee deep in the world of iPaaS trials and a HRIS integration strategy, where the real magic happens between vendor demonstrations and stakeholder expectations.

If a next-token predictor could decipher the labyrinthine requirements of enterprise integrations, figure out which custom workflows won’t immediately collapse under real-world usage, navigate the political minefield of backend architecture decisions, deliver cost-benefit analyses that acknowledge the existence of technical debt, sit through meetings where everyone talks past each other while pretending to agree, and handle the beautiful chaos of enterprise software politics, then at that point I’d start to get worried.

But I know what you’re all here for. You want to know how next-token predictors played a role in my contract termination.

So was I laid off because of AI?

Yes, but with a twist.

Fittingly, on my last day there was a company wide All Hands where one of the company leaders delivered a presentation on future direction. He elaborated on the recent decision to terminate the contracts of 20 developers and then went into a pitch:

“I think the industry is changing and we have to change with it”

“Historically, we’ve hired backend developers but we want to move towards hiring Builders. We’ve assigned job titles like ‘You are a Ruby developer’, and ‘You are a React developer’, but I think we’re now moving away from that. With AI tools we can now have more developers working across many different languages in a more generalist fashion. We are now going to be a culture of Builders”

His enthusiastic embrace of the term “Builder” suggested he’d missed the memo about Builder.ai’s spectacular failure who, instead of selling an advanced AI solution, turned out to be 700 Indians in a trench coat.

The conclusion to draw from this presentation was that if you’re a frontend developer within the company, you will now be expected to do more backend work, and vice versa if you’re a backend developer. In a way I think this is a noble idea, as I’ve always felt quite limited in my career to being pigeonholed as “just” a Ruby developer. I’ve always loved the idea of working on a team and getting my hands dirty in backend API code, frontend code and also being responsible for deployments.

But the company needs to be set up correctly for that. Mentorship and support needs to be built in, not just an after thought. Expectations on velocity need to be managed carefully. Stakeholders need to be aware that things will likely move more slowly as developers are given tasks way outside of their area of expertise and that next token predictors, for all their strengths, still have limitations.

This wasn’t a thoughtfully designed cross-training program, however, but rather a panic induced mathematical necessity disguised as professional development and forward thinking.

Reading between the buzzwords, the message was simple: specialists were now generalists by executive decree. Frontend developers would be conscripted into database duty, while backend engineers would learn to wrangle CSS frameworks.

The transformation of the engineering team from specialists to swiss-army-knife developers just happened to coincide with the departure of twenty colleagues. Twenty fewer salaries meant twenty more responsibilities distributed amongst the remaining staff. The survivors would now juggle unfamiliar codebases with the grace of circus performers but instead of applause they’d receive performance reviews questioning why release cycles had slowed and bugs had increased.

They’d now perform their new multi-disciplinary roles like technical acrobats, bending across unfamiliar frameworks while leadership boasted to investors about “AI-driven efficiency gains” (corporate speak for “we discovered how to make three developers do the work of five but for the same salary.”)

Meanwhile, the actual AI assistance would manifest like a comedy sketch, as frustrated backend engineers plead with ChatGPT: “For the love of all that is holy, I just need this div to center properly please write the CSS correctly! Fix it or you go to jail !”

I don’t want to paint these senior leaders as villains because to do so would ignore the institutional terror driving their thinking. While it’s tempting to dismiss these leadership choices as executive malpractice, there’s a human element worth considering. The truth is they’re likely just as anxiety laden as the rest of us and are coming up with short sighted decisions purely for self-preservation but dressed up as innovation. In their heads its miss the AI wave and become the next Blockbuster CEO or embrace it recklessly and maybe buy yourself another quarter of relevance.

A recent survey of CEOs revealed that three quarters of them say they could lose their jobs within two years if they don’t deliver measurable AI-driven business gains. They also anonymously reveal that one third of their AI initiatives are “just for show” (what the industry commonly refers to as “AI washing.”) The result is a feedback loop of increasingly desperate decisions, where laying off experienced staff to funnel money into projects that shove next-token predictors into every facet of your organisation feels safer than explaining to shareholders why you’re not riding the AI wave.

So did I lose my job to AI? Yes, I lost my job to AI, but the AI in question was the artificial intelligence of corporate decision-making, not the technological kind. It wasn’t because a next-token predictor replaced all the tasks I do in order to perform my job. It’s because corporate leaders have completely overdosed on AI marketing materials and can no longer distinguish between reality and vendor demonstrations and they’re also driven by the primal fear of professional extinction.

The damage done by the complete cognitive capture of leadership by AI evangelists will only be clear once the dust settles on this current trend. In the coming years we’re going to see some truly horrifying codebases, a coming apocalypse of spaghetti code, as the expectation from the C-suite to write code faster becomes more entrenched.

We’re in for a difficult few years, both as developers of software and users of software. Quality will become the luxury good nobody can afford, because in a market where everyone’s software is broken, the company that insists on working products suddenly looks expensive. We’ve managed to create a market so dysfunctional that dysfunction itself has become the only rational strategy.

Developers will need to adjust their quality standards to match corporate reality. Users will need to embrace enshittification on steroids as bugs get rebranded as “acceptable failures”. It’s a race to the bottom disguised as innovation, where the last company to lower their standards loses market share. Once broken software becomes the norm, fixing things properly starts looking like a competitive disadvantage.