The AI jobs numbers look calm until you cut them by age. This episode examines the June 2026 Stanford Digital Economy Lab and ADP Research data, which shows early-career workers in the most AI-exposed jobs contracting near 4 percent a year while their least-exposed peers keep growing, and it works through why the entry-level on-ramp is the first thing AI removes.

The conversation covers the augmentation versus automation split that decides whether jobs grow or vanish, the limit of a population dashboard that can diagnose the trend but cannot see whether any single worker is being grown or replaced, and the case for measuring the person in real time with a named human at the checkpoint. It closes on the hardest question the data raises: if AI absorbs the tasks that used to train people, where does the next generation of capable, accountable  ...