Independent New Hampshire Journalism · Northfield, NHWhat it said: That the author is “mounting a campaign for Governor of this state” and “running for Governor of New Hampshire,” under a section headed “What a Governor Should Say About This.” That 18 to 22 percent of “all online job postings in 2025” are ghost jobs, per Greenhouse. That 7.4 million June 2025 openings against 5.2 million hires means “2.2 million listings that vanished into the void in a single month.” That 93 percent of HR professionals are “running a soft fraud.” That “Fast Company’s recruiter survey” found 81 percent. That “a Clarify Capital study” found nearly one in three employers post listings they do not intend to fill. That the odds a posting is real are “in the most generous reading, about two-in-three.” That “New York City, Colorado, and Connecticut” prohibit graduation-date fields, “as three states already do.” That a CWI Labs survey of “1,600 adults over fifty” ran in August 2024. That DOGE-related layoffs “account for more than 280,000 job losses this year.”
What is true: Dexter Dow is not a candidate for governor and has never been a candidate for public office. Those sentences should not have run. The Greenhouse figure is an internal review of postings on Greenhouse’s own platform for the second quarter of 2024, not all online postings in 2025. BLS job openings are a point-in-time count on the last business day of the month and hires are a flow across the whole month, so the two cannot be subtracted; unfilled openings carry into the next month’s count. The LiveCareer survey covered 918 self-selected HR professionals, and the 48 percent who post occasionally described pipeline-building and seasonal hiring rather than fraud. The 81 percent figure is from a 2024 MyPerfectResume Recruiting Trends survey that Fast Company cited, and it records recruiters saying their employer posts such ads. The “three in ten” finding is ResumeBuilder.com’s May 2024 survey of 1,641 hiring managers, not Clarify Capital. On the article’s own figures, the most generous reading is that about 82 percent of postings are real and the least generous about 69 percent — “two-in-three” is the worst case, not the best. The states restricting graduation-date questions are Connecticut, Colorado, Oregon and Delaware; New York City is not among them and New York’s bill is pending. The CWI Labs report is dated September 24, 2024 and surveyed 1,600 Americans aged 18 to 71; the 59 percent figure applies to job seekers aged 50 and over. The 280,253 DOGE figure is announced layoff plans for federal workers and contractors over February and March 2025, not realised job losses across the year.
What changed: Every reference to a gubernatorial candidacy was removed, including the section heading. The statistics above were reattributed to their actual sources and rescoped. The openings-minus-hires inference and the “soft fraud” characterization of 93 percent of HR professionals were removed. The odds paragraph, the graduation-date states and the scorecard were corrected.
Still under review: The claim that corporate jobs in 2025 are down roughly 3.5 percent from 2021, the ResumeUp.AI LinkedIn figures, and the author’s own application and rejection counts carry no citation and could not be independently sourced. They are unchanged and should be read as unsourced.
When: September 19, 2026, 2:35 a.m. ET.
Five Thousand Nos
A field report from the collapsed corporate ladder — and the ghost economy that replaced it.
I just closed out another application cycle. A company with more than two hundred open positions. Not one of them for me. Add it to the pile.
Seven years. Five thousand-plus applications. One verdict.
I used to believe what we’re all told to believe: the market is a meritocracy, effort compounds, the cream rises. I don’t anymore. Not because I stopped trying — I never stopped trying — but because the evidence has piled up into something I can no longer pretend isn’t there.
Let me walk you through what the data says about the system that has been telling me “no” for the better part of a decade.
The Ghost Job Economy
Between 18 and 22 percent of the job postings on the hiring platform Greenhouse were ghost jobs — listings with no intent to hire — in the second quarter of 2024, according to Greenhouse’s own internal review of its client data. On LinkedIn, an analysis by ResumeUp.AI puts the figure at roughly 27 percent, and above 30 percent in Los Angeles.
The federal data does not settle the question either way, and it is worth being careful about why. In June 2025, employers reported 7.4 million openings to the Bureau of Labor Statistics and made 5.2 million hires. Those two numbers cannot simply be subtracted: openings are counted on the last business day of the month, hires across the whole month, and an opening that goes unfilled rolls into the next month’s count rather than disappearing. What the series does show is a persistent gap between how many roles employers say are open and how many people they actually put on payroll.
Here is the number that got me: in a March 2025 LiveCareer survey of 918 HR professionals, 45 percent said they “regularly” post ghost jobs, and another 48 percent said they post them occasionally — the second group mostly citing seasonal needs and talent-pipeline building. Even on the narrow reading, nearly half the respondents said they routinely advertise work they are not hiring for.
A 2024 Recruiting Trends survey by MyPerfectResume, cited by Fast Company, found 81 percent of recruiters saying their employer posts ads for jobs that either do not exist or are already filled. A ResumeBuilder.com survey of 1,641 hiring managers in May 2024 found three in ten companies had active fake listings up at the time.
Translate that for anyone still laboring under the illusion of a functioning labor market: when you spend a weekend polishing a resume, writing a tailored cover letter, and customizing your portfolio for a specific opening, roughly one posting in four or five is not a real job. In some cities the rate runs closer to one in three. And nobody tells you which is which.
That is the system. It is not broken. It is working exactly as designed for the people running it. The only ones it is broken for are the applicants — and we are treated as an externality.
The Algorithm Doesn’t Like Your Birthday
I am thirty-eight. Technically not yet old enough to invoke federal age-discrimination protections, which begin at forty. But the algorithms do not wait for my birthday.
AARP testified before the U.S. Senate Special Committee on Aging in September 2025 that algorithms used to scan resumes and applications “can accelerate bias using graduation dates or years of experience as proxies for age.” Two-thirds of workers over fifty report having seen or experienced age discrimination on the job. In a CWI Labs report released September 24, 2024, based on a survey of 1,600 Americans aged 18 to 71, 59 percent of job seekers aged 50 and over said their age had created obstacles during hiring.
Meanwhile, middle-management ranks are being hollowed out. Challenger, Gray & Christmas attributed more than 280,000 announced layoff plans among federal workers and contractors to DOGE actions across February and March 2025 alone. The people flooding back into the market are mid-career professionals — the exact demographic the algorithm is trained to flag as expensive, inflexible, or “overqualified.”
States are beginning to notice. Connecticut, Colorado, Oregon and Delaware now restrict employers from asking for graduation dates or other age-revealing dates on job applications, for this exact reason: a graduation date is not a qualification. It is an age calculator. And employers have been using it to do what the Age Discrimination in Employment Act forbids them from doing in person.
I am not old. The algorithm and I are, nonetheless, not on speaking terms. And the algorithm is doing the hiring.
The “Didn’t Try Hard Enough” Lie
I am, among other things, a workaholic. I do an enormous amount of work for free, for the satisfaction of organizing projects and building things of use. I write books. I run a civic journalism publication. I build software. I produce video. I write daily. I ship.
None of it, apparently, counts on a resume screener.
The standard-issue lecture from a certain generational cohort — the one that bought houses on a single income at twenty-four — is that people like me did not grind hard enough, did not network hard enough, did not want it badly enough. The lecture would be insulting if it were not so confused. The ladder they climbed no longer exists. It is not that the rungs are farther apart. Somebody sold the ladder for scrap and is now lecturing the rest of us for not climbing.
I have been maintaining a file. Five thousand rejections. Auto-replies. Silent ghostings. Templated regrets from recruiters who never opened the application. “We have decided to pursue other candidates” for jobs that were never real. “We will keep your resume on file” in a file that does not exist.
I am thinking of publishing it. Bound, indexed, cross-referenced. A monument to the lie.
| The Premise You Are Supposed to Accept | What the Surveys Say |
|---|---|
| The job you’re applying to is a real, open role. | ~18–27% are not |
| HR professionals don’t post listings they have no intent to fill. | 45% say regularly |
| Recruiters’ employers don’t advertise filled or fictional jobs. | 81% say they do |
| Age doesn’t factor into hiring until the legal threshold of 40. | Algorithms filter earlier |
| Companies don’t keep fake listings live. | 3 in 10 currently do |
| The problem is that applicants aren’t trying hard enough. | See above |
What This State Should Do About It
Let me say the thing most people running for most offices will not say, because the donor class does not want to hear it:
The American labor market is, in significant part, a simulation. The employment figures are distorted by postings nobody intends to fill. The “skills gap” is frequently a euphemism for “we will not pay for the skills we already have.” The gig economy is a pay cut in yoga-studio vocabulary. The algorithm is discriminating on age at scale and the state is calling it efficiency.
New Hampshire has the tools to lead on this, and they are not radical tools. The state could require employers advertising jobs here to disclose intent to hire, a salary range, and how long a position has been recycled. It could prohibit graduation-date fields on applications for state-licensed or state-regulated employers, as Connecticut, Colorado, Oregon and Delaware already do in various forms. It could put actual enforcement behind the age-discrimination statute at the hiring stage, not just once somebody is already on payroll. It could stop cheerleading “job postings” as a proxy for economic health when a meaningful share of them are fiction.
None of that gets done while we keep pretending the system works and the problem is the applicant.
The Punchline
The bitter joke in all of this — and I do find it funny, in the sense that a coyote finds the trap funny after the third time — is that the people who actually know me know I am a workaholic. I produce constantly. I ship constantly. I give work away constantly, for the satisfaction of building something of use to somebody.
Would I hire me at this point? Honestly, no. I would be a pain. I would be critical of everything. I would refuse to pretend the meetings were productive or that the strategy deck made sense or that the CEO was as sharp as his salary implied. I would be a disaster of a hire for a company running on ceremony.
So the algorithm and I have reached a quiet understanding. It does not want me. I do not want it.
What is left is to do the work anyway — for the state, for the people getting the same templated rejection I am, for the next person about to spend a Saturday polishing an application for a job that was filled in April or never existed at all.
Five thousand nos. Monument pending.
Corrections: Granite State Report corrects verified errors promptly and appends a dated note identifying what changed and when. The corrections policy and the corrections register are public, and correction notes are never removed.

