- Offboard
- Posts
- The U.S. Just “Lost” 911,000 Jobs
The U.S. Just “Lost” 911,000 Jobs
GenAI rewires the how • More cuts at Scale, Microsoft, Rivian • BLS trims 911,000 jobs • Tools: FiveShots + Huntr
"When your primary goal is to be liked, you can't take risks. You can't disagree. You can't push boundaries. You become a prisoner of other people's expectations." -FS👋 Welcome, Jobseekers
GenAI is rewriting how work gets done, not erasing people. At the same time we’re seeing fresh cuts at Scale AI, Microsoft, Rivian and more, and a BLS preview just trimmed about 911,000 jobs from last year’s total. Translation: hiring will get choosier and faster-to-value. Need speed? Use FiveShots to lock a pro headshot and Huntr to sprint applications.
🧑✈️ Career CoPilot: AI at Work Report: GenAI Will Rewrite the How, Not Erase the Who
📉 Layoff Report: Scale AI, Microsoft, Rivian All With More Cuts
📈 Trends & Data: The U.S. Just “Lost” 911,000 Jobs
🔧 Jobseeker Tools: FiveShots, Huntr
Scale AI, Microsoft, Rivian all with more cuts
Source: TrueUp
Scale AI Makes Cuts to Key Team After Meta's $14 Billion Investment
Microsoft cuts 42 more jobs in Redmond, continuing layoffs amid AI spending boom
Scope3 Lays Off Staff as Part of Shift Toward Agentic Advertising
EV Maker Rivian Lays Off Workers as It Preps Launch of Cheaper SUV
Bobble AI lays off 50 employees amid organizational redesign
AI startup theGist shuts down, plans to return investor funds
Indeed’s AI at Work Report: GenAI Will Rewrite the How, Not Erase the Who
Overview
Indeed just released its AI at Work Report 2025: How GenAI is Rewiring the DNA of Jobs and they found that almost half of the skills in a typical job can now be executed by GenAI with humans supervising, with software development most exposed and nursing’s core work largely insulated.
TL;DR
What they did: ran GPT-4.1 and Claude Sonnet 4 across 2,884 real skills, 15 times each at temperature 0.
Top numbers: a typical job is 46% hybrid/full, 12% assisted, and 42% minimal. Only ~1% of skills look fully automatable now. 19 skills (0.7%) flagged as very likely to be replaced.
Quick play: measure the model on your tasks, then show how you fix it.
The Details
Indeed started with 2,884 distinct skills pulled from job data across roles and industries. The team fed each skill into two model families, ran each model 15 times at temperature 0 to reduce randomness, and combined the outputs into a repeatable pipeline. That process produced four labels for every skill: minimal, assisted, hybrid/full, or fully transformable. The skill labels were rolled up to compute job-level exposure.
The headline numbers come straight from that pipeline. On average, 46 percent of a job’s listed skills were marked hybrid/full, 12 percent assisted, and 42 percent minimal. The same method produced sector splits: software development clustered high on hybrid exposure, while nursing skewed toward minimal. The authors flagged limits: only about 1 percent of skills appear fully transformable today, and 19 skills are near-certain candidates for replacement.
Why It Matters
These are measured outputs, not opinions. That makes the 46 percent figure a snapshot of current model capability, not a prediction of mass layoffs. The multi-model, multi-run design raises confidence in the middle of the distribution. It also exposes where models still fail: edge cases, judgment calls, and quality control.
Treat the report as a capability map, not a timetable. Employer adoption, policy, and hiring choices will decide how fast roles actually change. The practical opportunity is clear: document and prove the specific skills where you add judgment, tests, or fixes. That is the kind of proof hiring teams can evaluate in interviews and take-home tests.
Sources: Indeed Hiring Lab, AI at Work report, Annina Hering and Arcenis Rojas, September 2025.
The U.S. Just “Lost” 911,000 Jobs
Overview
The BLS preview says the labor market was softer than we thought. If your search felt sluggish, you weren’t imagining it.
The Details
Preliminary benchmark cut: -911,000 jobs (-0.6%) to total non-farm for the 12 months ending March 2025. Final revisions arrive Feb 2026.
That implies ~71,000 average monthly gains, not ~147,000.
August snapshot: +22,000 payrolls; unemployment 4.3%.
Markets price a near-certain rate cut on Sept 17; small chance of 50 bps.
Last year’s prelim mark-down was -818,000; the final cut landed at -598,000.
Street expects August CPI ~2.9% YoY on Thursday; a hotter print could limit a bigger cut.
Why It Matters
The BLS revision means the labor market is weaker than many thought. Fewer net hires equals more candidates chasing each open role. Companies will respond by tightening budgets, slowing decisions, and favoring hires who show immediate, measurable value. That raises the bar for every stage of the job search: sourcing, screens, interviews, and offers. A Fed rate cut might ease funding and loosen budgets down the road, but that effect takes time. In the near term, expect hiring to be choosier and slower.
Career implications
Expect tighter screens and slower recruiter outreach. Recruiters will use stricter filters and automated screens. You’ll see longer turnarounds, more interview rounds, and slower feedback. Move faster with follow-ups and concise evidence of impact.
Roles tied to profitable growth and cost control will move faster. Hiring will favor product, sales, ops, and engineering work that directly increases revenue or cuts expenses. Lead with numbers: revenue gained, cost saved, retention improved.
Short-term and flexible work will spike. Companies will prefer contractors, temp-to-perm, and internal backfills to limit risk. Consider contract gigs, consulting, or internal moves as faster routes to pay and runway while full-time openings tighten.
FiveShots
From selfie to pro headshot in minutes
Upload one selfie. Get five LinkedIn‑ready portraits you can use anywhere in minutes. Professionally polished. For $5.
Huntr
Huntr helps you create tailored resumes and cover letters fast with AI, fill out application forms in one click, and automatically organize your job search.
Reply