A paper: O-Ring Automation.

Makes the case that estimates of labor displacement due to automation are structurally too high due to how labor is offset in real workflows.