The pattern is the same everywhere you look: companies in the black cutting real engineers and routing the savings into AI infrastructure. May 20 was the single worst day of 2026 — roughly 11,000 jobs announced between Meta and Intuit alone. The cumulative number, 142,000 across more than 212 companies in under five months, is not a correction. It's a structural shift. The hiring picture splits cleanly along a fault line. AI/ML, model evaluation, applied research, and AI safety are in acute shortage — companies burning through engineers on one side are quietly competing for a much smaller pool on the other. Traditional software engineering, product management, and recruiting are contracting. A Resume.org survey of 1,000 US hiring managers found 55% expecting further cuts, with 44% naming AI as the primary driver. If you're advising a team on headcount planning right now, those numbers are the starting point for the conversation.