Reporting archive
Evidence on AI and work
-
United States Tech Force opens 1,000 two‑year AI roles
Washington just turned AI hiring into policy—a 1,000‑role, two‑year federal tour that could reset where ambitious engineers start and what industry values.
-
Permits, liability, and the moat LLMs can’t cross
As models devour desk work, the safest way to stay indispensable is to hold permission—get a state license and let AI handle everything around the edges.
-
120,000 tech layoffs bankroll GPUs and cloud capacity
2025’s tech cuts weren’t austerity—they rewired org charts into pipelines, thinning the middle and pouring budget into GPUs, cloud, and security.
-
Washington’s AI rollback picks substitution as workers lose the ladder
AI’s big week wasn’t a bubble story—it was the moment we optimized away entry‑level work while policy waved it through.
-
Congress wants receipts on AI layoffs by December 31
Congress just started the stopwatch on AI-attributed layoffs, demanding company-by-company receipts on who was cut, how, and why by year-end.
-
Vanguard’s 43% task time shave meets Howard Marks’s labor math
Howard Marks’s postscript reframes AI as a credit problem: margins rise, ladders vanish, and the social floor that keeps coupons safe starts to crack.
-
Big banks publish the AI timeline for white‑collar reductions
Wall Street’s biggest banks just moved AI from demos to org charts, tying measured productivity gains to fewer backfills and leaner operations.
-
Tariffs, not transformers, are chilling entry-level hiring
The youth hiring slump looks less like an AI cull and more like a classic margin-driven freeze—tariffs, policy whiplash, and CFOs stretching teams—so watch vacancies and quits, not the hype.
-
Jamie Dimon signals 2026 isn’t the AI layoff year
Jamie Dimon recasts AI’s job threat as a tempo issue—urging phased rollout, regulation as metronome, and short-run job gains instead of a panic purge.
-
Microsoft maps 200,000 Copilot prompts to U.S. jobs
Microsoft turned 200,000 real Copilot chats into a live map of which tasks AI already does—and how that’s quietly rewiring roles, review gates, and spans of control.