Jobs AI will eliminate, in order (Musk, Jensen Huang and 12 others)

An independent exercise, not a statement by the video creator or verification of every claim in the video.

Is an 11% decline the same as a relative gap?

This hypothetical comparison does not reproduce an actual research estimate. Start each group at employment of 100, then let the highly exposed group reach 89 and the comparison group 110. The exposed group declines by 11% from its own baseline, while the comparison group grows by 10%. The final level difference is 21, and the gap relative to the comparison group is 21 ÷ 110 × 100 ≈ 19.1%. Calling these different calculations the same decline rate can overstate or understate a result. In an actual report, check the baseline period, group composition, indexing and controls before treating a gap as an AI causal effect.

Check for yourself

1. Does the reported decline compare a group with its own past or with another group’s trend?
2. Does the outcome measure employment, hiring, exits or unemployment?
3. What separate evidence shows whether an AI assistant helps novices build skills or reduces new hiring?

Limits of application

100, 89, 110, 11%, 10%, 21 and 21 ÷ 110 × 100 ≈ 19.1% are hypothetical and do not reconstruct the Stanford estimator. The comparison does not control for composition, the business cycle, education or company characteristics and establishes no causal effect.