If machines do most of the work, what do the humans do?: Institutional reform
New to public policy
The Luddites were not wrong
On March 11, 1811, textile workers in Nottinghamshire smashed the stocking frames taking their jobs. The name became a slur. What gets left out: hand-loom weavers went from decent living in 1800 to destitution by 1830. A generation was destroyed. The long-term gains were real. The long term was fifty years away.
Every workforce transition follows the same arc: technology faster than institutions, a generation bearing the cost, the next inheriting gains. The AI wave is faster but not categorically different.
The GI Bill, not the check
The UBI advocates want to write a check. A check is a confession that the institutions gave up. The GI Bill did not write checks. It opened doors — tuition, training, home loans. Eleven million veterans entered the postwar economy, and the institutions channeled them so effectively the investment paid for itself in a decade.
The equivalent today is not retraining as it exists. It is redesign: German-style apprenticeships, community colleges retooled for care work and skilled trades, portable benefits decoupled from employers. The meaning-crisis camp says the problem is deeper. We note that the deepest meaning crises — the opioid epidemic, deaths of despair — correlate precisely with institutional failure: counties where manufacturing collapsed and nothing replaced it.
A person does not find a vocation by introspection. She finds it by being embedded in a community that needs something she can provide, in a structure that trains her.
Where we concede ground: Our track record is worse than our rhetoric. Decades of programs, billions spent, negligible earnings gains.
What would change our mind: If a comprehensive institutional framework produced outcomes no better than a simple cash transfer of equivalent cost.
Read the full synthesis: If machines do most of the work, what do the humans do?