"Already Obsolete": How Software Engineers Actually Talk About AI Dread

There's a specific whiplash that mid-career engineers keep describing, and it goes like this: by day, the AI assistant makes you feel five times as productive — genuinely thrilling, the best tooling of your career. By night, the exact same capability reads as a countdown. I'm training my replacement. Same tool, same person, euphoria or dread depending on the hour.

That oscillation is the honest texture of AI career anxiety, and it's nothing like the headlines. In the forums where engineers talk to each other anonymously, the fear is rarely about a specific model release. One developer put it in a sentence that's been quoted across the industry: "I used to feel confident. Now every day I wonder if I'm already obsolete." Not an event — a hum. The dread is ambient and narrative-driven, and that turns out to matter enormously for what actually helps.

The hum, not the event

Ask senior engineers — ten, fifteen, twenty years in — and a surprising pattern emerges: most of them aren't afraid of the tooling. They use it daily and know its limits better than anyone writing headlines about it. One senior dev mapped his own arc in a heavily upvoted thread: a full range of emotions since 2022, "from total doom and gloom" to genuine skepticism — and I say this as someone who uses them daily. The tools themselves were the cure; daily use calibrated the fear. What's grinding seniors down instead is the narrative: the drip of CEO statements, LinkedIn prophecy, and benchmark headlines declaring their profession over. The forums answer it with dark humor — one dev set a literal calendar reminder for a lab CEO's "90% of code in six months" deadline ("this mania got out of hand"); another's line got a wall of upvotes: "I'll worry about AI doom when companies switch to LLM bookkeeping."

The sharpest distinction in those threads is one no headline makes. When a non-technical product manager ranted at a dev team that they'd all be replaced soon — citing a CEO clip, treating AI, as the dev put it, "as some God" — the sober top comment cut to the real fear: the danger isn't that AI can replace engineers. It's that executives will act as if it can, eat the losses later, and in the meantime heaps of people lose jobs for no reason. Fear of the model and fear of decisions made in the model's name are different fears. Only one of them is about technology.

There's a second, quieter cost that one developer named exactly: "slow, invisible burnout" — remote isolation, AI uncertainty, and imposter syndrome compounding into something nobody's standup covers. And its engine is the treadmill: "I feel like I'm burning out, but I can't stop to breathe. If I do, I'll fall behind." Read that twice — the keep-up compulsion, not the robot, is what's actually consuming people.

The safe-path kids got it worst

The anxiety is not evenly distributed. Seniors report fatigue; juniors and students report despair — and the market data says their asymmetric fear isn't irrational. New-grad CS unemployment has been running around six percent — roughly double philosophy majors, a punchline nobody in the field finds funny — with entry-level postings down by nearly a third year over year. The student forums have ritualized it into two words that answer every post: it's over.

Underneath the meme is an identity injury that the posts keep circling: these are the people who chose computer science because it was the safe path — the smart kid's rational choice, often with family expectations riding on it. The spiral isn't just "will I get a job." It's I was the one with the plan, and the plan dissolved. When your career was 90% of your self-concept, every model release threatens 90% of you — which is why the identity work matters as much as the job search.

And the press supplies the nightmare made flesh: the widely covered story of a veteran engineer, twenty years in, who lost a $150K job in an AI-attributed layoff, sent eight hundred applications, and ended up gig-driving and living in a trailer — writing that "the great displacement is already well underway." Worth noting: the forums fought bitterly over that story, half seeing themselves in it, half blaming his specific choices. Both camps refreshed the layoff trackers a little more often that week.

The doomscroll is a self-harm loop with a research alibi

The junior posts have a signature confession: checking the AGI-timeline subreddits nightly, insomnia, and an inability to focus at work because what's the point if it's all automated in three years? The top-voted responses in those threads are consistent and unsentimental: log off — the doomer subs are engagement farms. And the best one: build something with the tools, so the fear becomes information.

That last line deserves a job title. Doomscrolling feels like due diligence — you're "staying informed" about your own disruption. But you can't research your way to certainty about 2030, and every refresh is another rep for the overthinking loop. One dev hedged by secretly working two remote jobs and reported the hedge itself became the second fear — terrified of being caught, still wondering about AI before every 1:1. Anxiety compounds faster than income.

And here's the pattern the archives make undeniable: AI dread never travels alone. In thread after thread it arrives fused to something older — a ten-year veteran's post that's ostensibly about AI is really about feeling he'd had "the same year of experience over and over" (imposter syndrome wearing a new mask); a contractor's AI fear is stacked on offshoring, a layoff, and skill stagnation; a senior's spiral is amplified by a growing family and the question underneath every benchmark headline: how do we provide for our families if we automate our careers out of existence? Same headlines, wildly different dread — the variable is almost never the model. It's the runway, the obligations, and the older fear the AI narrative gave a new name to.

What the upvoted advice actually says

Strip the noise and the forums' own consensus is better than most careers content. Use the tools daily — familiarity converts fear into a map of what they can't do. Build runway — "live well below your means in a volatile industry" is the advice that keeps getting upvoted, because money converts an existential threat into a solvable timeline. Bet on judgment, not syntax — the job was never typing code. Quit the doomer feeds. The honest middle position the community keeps landing on: the job will change and juniors will genuinely have it harder — "same number of people, but not the same people" — but replacement-by-next-year talk is investor theater. And the forums police the line themselves: posts with a specific situation and a specific question get engaged, thoughtful help; abstract headline-driven doom gets mocked or removed. What gets ridiculed alongside it: "just learn a trade," "pivot to prompt engineering," "just do a side project" (resented as generic), and pure denial. The community respects neither doom nor cope. It respects accuracy.

One more finding worth naming, because it's a gap: in all those threads, explicit mentions of therapy or any mental-health practice are strikingly rare. The closest anyone gets is a quiet "everyone needs mental care" buried mid-thread. Engineers self-treat with information, money, and humor — all three genuinely useful, none of them built for the 2am version of the problem.

Which is the whole opportunity, mental-health-wise: this anxiety responds unusually well to honest reasoning. "I'll be obsolete in two years" is a claim, not a fact — and engineers, of all people, know how to debug a claim. What's in the evidence, what's extrapolation, what's 3am catastrophizing wearing a data costume? What's actually in your control this week — what you build, learn, save, and say no to — and what belongs to the frontier labs' release schedule? That's exactly the conversation ILTY is built to have with engineers: a companion that pushes back on your reasoning instead of agreeing with your doom, tracks the superhuman/doomed oscillation until you can see it's input-driven, and keeps dragging the conversation back to the controllables. Yes — an AI companion for AI dread. Use the tools. That was always the job.

Frequently asked questions

Is it rational to be anxious about AI taking my engineering job? The concern is rational — the market data for entry-level roles is genuinely bad, and the industry is repricing. The 2am spiral is not: it treats an uncertain multi-year trajectory as a settled verdict. The skill is holding both — take the trend seriously, refuse the certainty.

How do I stop doomscrolling AI news? Treat it as exposure, not information: time-box it (one deliberate session a week beats thirty ambient ones), mute the timeline-prophecy subs, and replace the checking habit with building — the forums' most durable advice, because using the tools daily converts dread into an accurate map.

Should I leave software engineering? Maybe — but decide from a plan, not from a panic. The people who navigate this well cite the same sequence: runway first (it buys thinking time), then honest inventory of what your judgment is worth beyond syntax, then deliberate moves. Fleeing to a random "safe" field at the bottom of a spiral is how people trade one anxiety for two.

When is AI career anxiety something more serious? When it stops being episodic: insomnia most nights, panic attacks, an inability to work at all, or a flat dread that doesn't lift when the news cycle does. That's territory for a professional, not a subreddit — and ILTY will say exactly that if it hears it.

ILTY is an honest AI companion for thinking clearly while the ground moves — one that pushes back on your doom instead of agreeing with it. See how it fits an engineer's head, or try it free.