Specialists: From Barber-Surgeons to Domain-Specific AI Swarms

In 1720 London, Thomas Rawlings operated from a single shop on Fleet Street. With one razor, he would shave your beard, pull your rotted teeth, lance your boils, and amputate your gangrenous limbs. The red and white barber pole still spinning outside modern shops? That's blood and bandages—the universal answer to every question a human body might ask.
This is the story of how that world ended. How humanity discovered that complexity exceeded what any individual could master. And how our response—specialization—is now reshaping artificial intelligence itself.
We begin in 1540 when Henry VIII signed a charter binding barbers and surgeons together. We witness the 1745 divorce when surgeons declared independence. We meet John Hunter, the illiterate Scottish "dunce" who transformed surgery from craft to science with 14,000 specimens. We follow Adam Smith into a pin factory where ten workers produced 48,000 pins per day—240 times more than they could make individually.
We stand on Ford's assembly line. We enter Johns Hopkins where Halsted created the residency system. We arrive at Mission Control on July 20, 1969, where 400,000 specialists coordinated in real time to land humans on the Moon—and a 26-year-old guidance officer's split-second decision, based on specialized knowledge no generalist could possess, saved the mission.
Then we enter Marvin Minsky's office in 1985, where his "Society of Mind" proposed something unsettling: there is no unified intelligence. The mind is a collection of specialized agents, each mindless by itself, producing intelligence through coordination. "You can build a mind from many little parts, each mindless by itself."
The AI community ignored him for forty years—until 2023, when Stanford researchers discovered that large language models suffer from "Lost in the Middle": they forget information in the middle of long contexts. The promise of one giant model doing everything was hitting the same wall that broke the barber-surgeon. The solution? Multi-agent AI systems: research agents, analysis agents, writing agents, fact-checking agents—coordinated specialists achieving what single models cannot.
"No one knows everything; together, they know enough."
From Thomas Rawlings' single razor to AI agent swarms, from Adam Smith's pin factory to Mission Control to microservices architecture, we trace humanity's greatest organizational insight—and watch artificial intelligence rediscover it in real time.