Reporting archive
Evidence on AI and work
Reporting archive

Meta has set a new exchange rate—people for compute—and the first installment is due May 20.

Google just moved the interview to where the work lives: collaborate with Gemini, prove your prompts, and show you can keep AI‑accelerated code safe and fast.

Cloudflare just made agents the architecture—cutting ~1,100 roles, front‑loading rich severance, and daring the market to wait while it rewires the company.

Forget mass unemployment: a16z’s bet is on expansion and friction, while a 16% hit to AI-exposed new grads shows where the collateral damage starts.

Executives have found the perfect layoff alibi—AI—while the missing metrics tell a different story.

Coinbase is melting its org chart into software—cutting 14%, flattening to five layers, and betting one person plus agents can outship yesterday’s teams.

A Hangzhou court just ruled that “AI made us do it” isn’t a legal excuse—if you automate, you must negotiate, retrain, or compensate.

AI’s productivity surge won’t translate into a boom unless the gains flow beyond profits to the paychecks that keep demand alive.

AI’s job story is being written in copper and concrete as unions and tech quietly team up to build the megawatts—and manpower—the models need.

Singapore recasts AI adoption as labor-market choreography—hardwiring councils, incentives, and tool access to a single metric: time to reemployment.