Reporting archive
Evidence on AI and work
Reporting archive

The IMF just put numbers to the AI jobs debate: broad new skills grow headcount and pay, while AI-specific hiring thins the ranks—and the five-year timer is already running.

Meta just redirected a thousand-plus jobs from world‑building in VR to AI that can live on your face and in your hand.

When software roles become the biggest line item on London’s finance job boards, the org chart—and the power—shift to those who build and govern the machines.

AI’s labor shock is showing up at the door, not the balance sheet: fewer young entrants into exposed roles, unchanged exits, and a gate that won’t open.

This year’s AI bottleneck isn’t silicon—it’s the people who can orchestrate agents into reliable, measured throughput, and they’re about to reshuffle hiring, metrics, and power.

Angi just put “AI‑driven efficiency” into its 8‑K and 350 roles off its payroll, turning automation from a pilot into policy.

AI is boosting throughput while hollowing out the feeling of mastery—unless leaders rewrite roles, metrics, and status.

If AI is really driving the layoffs, where’s the productivity spike—and why do the announcements read like cover for mundane belt‑tightening?

What if AI nails the supply side but severs the paycheck pipeline that makes growth real?

WEF’s net job gains hide a harsher truth: the entry ramp is narrowing, signals are inflating, and the only number that matters now is the conversion rate from roles created to roles filled.