She was a video editor in Los Angeles — fifteen years cutting trailers, the kind of craftsperson whose name flashes by in three seconds at the end of something you loved. When the work dried up in 2023, she told herself it was the strikes. When it didn’t come back in 2024, she told herself it was the economy. This summer, she started driving for a delivery app between freelance gigs that never quite materialize. She is not a statistic. She is one of 370,000.
That’s how many jobs the U.S. information sector has lost since peaking in 2022 — a 12% decline over four years, across movies, music, telecom, media, data processing, and web hosting, according to Bureau of Labor Statistics data via FRED. In August alone, the sector shed another 23,000 jobs. And the remarkable thing is how little noise any of it made, because the August employment report was otherwise strong enough to let everyone look the other way.
It’s not enough to just read the headline jobs number — we must listen to the sectors quietly bleeding underneath it, learn what their decline is really telling us, and contribute to an honest accounting of who this economy is working for. Because 370,000 missing jobs is not a rounding error. It’s a city the size of Cleveland, emptied out one layoff notice at a time.
The sector the recovery left behind
The information sector is a strange statistical bucket. It holds the Hollywood studio and the telecom call center, the local newspaper and the data-processing warehouse, the record label and the web host. What unites them is that they all traffic in information — and all of them have been reshaped, ruthlessly, by the same forces: streaming economics that never quite replaced the old models, AI tools that do in minutes what teams once did in weeks, consolidation that turns three employers into one, and a advertising market that rewards platforms over publishers.
Twelve percent. Sit with that number. If your town lost 12% of its teachers, it would be a five-alarm civic emergency. But because these losses are scattered — a few hundred at a studio here, a thousand at a telecom there, a quietly shuttered regional paper somewhere else — they never arrive as a single event. They arrive as 370,000 private griefs, each one explained away as an individual story rather than a structural one.
And August’s 23,000-job drop landed in a month the broader report called strong. That’s the cruelest part: when the national number looks good, the suffering sectors become invisible. Nobody holds a press conference for the jobs that vanished inside a good report.
The paradox: software is booming next door
Here’s what makes this story genuinely strange rather than simply sad. While the information sector bled, software stocks staged a roaring comeback. The State Street Software and Services ETF — roughly 130 software and IT stocks — rose 5.2% on Thursday to a new all-time high, as investors concluded that AI won’t obliterate the software sector after all. Nvidia reported blowout earnings. The AI trade is alive and furiously well.
So we have two truths at once: the companies building the AI future are worth more than ever, and the workers in the industries being transformed by it are worth less — or at least, fewer of them are needed. The wealth is concentrating in the tools; the jobs are evaporating in the trades the tools touch.
This is the part where the conversation usually splits into the AI-doom camp and the AI-boom camp, and both miss the human middle. Consider one genuinely hopeful data point from the same week: a new Gusto report found that very small businesses that adopt AI actually grow employee headcount faster than those that don’t. The neighborhood shop that uses AI to handle its books and marketing doesn’t fire people — it hires, because the owner finally has time to grow. The job apocalypse, at least at the smallest scale, isn’t happening. What’s happening is a reallocation: routine information work consolidates into software, while human judgment, craft, and customer-facing work hold or grow.
My take: The 370,000-job decline isn’t primarily a story about AI replacing workers — it’s a story about business models collapsing faster than new ones form. Streaming gutted the economics of filmed entertainment. Private equity and consolidation gutted local media. Telecom stopped growing. AI accelerated all of it, but the rot in the old models predates the current boom. That matters because it changes the fix: you can’t retrain your way out of a broken business model, but you can build new ones — and the Gusto data suggests small, nimble businesses are doing exactly that.
Where the jobs are actually going
The Bureau of Labor Statistics released its long-term projections this week too, and they point the way forward: health care and social assistance are projected to see the most job growth over the next decade, driven by an aging population, even as overall job growth is projected to slow. The economy of the next ten years will be built less on content and more on care.
There’s something poignant and hopeful in that pivot. The video editor driving deliveries, the laid-off telecom technician, the newspaper reporter filing freelance pieces — the skills they carry (craft, communication, technical fluency, the ability to learn complex systems) are profoundly transferable. An aging country needs people who can show up, pay attention, and do careful work. Those are human capabilities, not information-sector ones.
But let’s be honest about the transition cost. A 50-year-old editor doesn’t become a nurse overnight, and “learn to code” was always a glib answer to a serious question. Retraining takes money, time, and a safety net — exactly the things disappearing jobs take away. If we take the 370,000 seriously, the policy conversation has to be about bridges: wage insurance, portable benefits, genuinely funded retraining, and an honest reckoning with the fact that “the economy is strong” can be true and inadequate at the same time.
Don’t let a good report hide a hard truth
August’s employment report will be remembered as a strong one, and in aggregate, it was. But aggregates are where individual stories go to disappear. The information sector’s four-year, 12% decline is the kind of slow-motion crisis that never triggers an emergency response — no single month is catastrophic enough, no single company large enough, no single headline dramatic enough.
So consider this article a small act of refusal: a refusal to let 370,000 people be a footnote. Behind that number are editors and engineers, reporters and riggers, producers and processors — people who built the movies we love, the networks we rely on, the media that informs us. Their work mattered. Their absence should matter too.
And here’s the hopeful part, the part worth carrying into the weekend: economies do reinvent their workforces. They have before, and the projections say care, craft, and human judgment are where the growth lives now. It’s not enough to just note the jobs are gone — we must listen to what the displaced know, learn from where the growth is actually heading, and contribute to building the bridges between the two. The next 370,000 jobs won’t look like the last ones. Our job is to make sure the people do.
