Interviewing is broken. Not "a little rough" broken — structurally, quietly, dangerously broken. And most of the people affected don't have the language for what happened to them until it's already over.
Here's what happened to me. I applied for a role. Had a great first conversation. Spent days going back and forth, genuinely excited. Then, out of nowhere: "Oh no — your resume caused a lockout. Luckily, I know someone who can help you reformat it." No red flags. Nothing I could point to in the moment. Just a very convincing, very patient operation designed to extract information from someone who was excited about a job.
I wasn't talking to a hiring manager. I was talking to a bot, or a scammer using one, running a script polished enough to survive days of real conversation.
Companies are flooded with fake candidates. Candidates are flooded with fake companies. Both sides are burning enormous time and energy just trying to figure out who's real — before either side has even gotten to the actual work of evaluating fit, skill, or need.
The "Two Developers, One AI" Fantasy
Layered on top of this is a newer myth: that two engineers with the right AI tools can now do the work of an entire department. Even in the best case, this isn't efficiency — it's a countdown to failure. Because a "team" was never just a headcount number to hit. It was distributed judgment.
Take away the team and you take away:
- Direction and vision — the managers, directors, VPs, and CTOs who decide what to build and why
- Quality — the QA, SDETs, and test engineers who catch what the builder can't see in their own work
- Security — the security engineers, DevSecOps, and CSOs who keep the thing from becoming a liability
- Infrastructure that survives contact with real users — DevOps and infrastructure engineers
- Design that actual humans can use — UX and UI designers
- Insight instead of guesswork — data engineers, analysts, and architects
- A structure that can hold weight — software architects
- A pipeline for the next generation — junior developers who become senior ones
Cut all of that and what's left isn't lean. It's undirected, unverified, unsecured, and blind to its own users.
Why This Isn't Abstract
It's easy to talk about "efficiency" until you remember what these systems actually touch: your mother's medical care. Your brother's insurance claim. The tools your kids learn from. The satellites overhead. The autonomous vehicle sharing a lane with an ambulance.
Running an organization without the people who maintain, verify, and question the work is a lot like running a railroad with no maintenance crews and no engineers. It's cheaper. It works — until it doesn't. And when it doesn't, someone else pays the cleanup cost: a family, a patient, a town whose backyard the wreck landed in.
The December Pattern
None of this happens in a vacuum. Layoffs cluster at year-end with a rhythm that's hard to ignore — a few percent one year, more the next, more again the year after — timed to land right before the moment revenue-per-employee gets bragged about in a room the laid-off employees will never see. Meanwhile the people actually doing the work — many with a decade or two of experience — cycle between unemployed, underemployed, and "next in line."
Quality Isn't Optional. It's the Whole Point.
There's no clean fix for information theft dressed up as a job offer, or for the incentive structure that rewards headcount cuts over reliability. But there is a starting principle worth saying plainly: quality matters, and how you build things matters.
That's the case for hiring real engineering teams — not a skeleton crew leaning on AI to fake competence, and not a headcount number optimized for a slide at the end of the year. The best and brightest, working together, is still the only reliable way to get something built right, kept secure, and made to last.