Instead of one processor grinding away step by step, thousands of processors attack different pieces of the problem at the same time and then combine their results. Modern graphics processors and AI chips still work that way today.
Later it occurred to me that civilization itself runs in parallel.
Billions of people are making decisions simultaneously—growing food, building machines, writing software, shipping goods, inventing ideas. No single person understands the whole system, yet somehow the pieces keep interacting and adjusting.
Two centuries ago those interactions were slow and local. A drought in one region could mean famine because food had no easy way to move quickly. Information traveled no faster than a horse.
Then the connections multiplied.
Railroads. Telegraphs. Radio. Container shipping. Satellites. Fiber-optic cables.
Each technology added new links between people and places. Over time those links turned the planet into a dense web of exchange.
Large networks behave in recognizable ways. Trouble in one place can spread quickly—financial crises, supply disruptions, pandemics. But the same connections that allow problems to travel also create resilience.
When crops fail in one region, food arrives from somewhere else. When a factory shuts down, production shifts to another country. The network bends, reroutes, and adjusts.
In other words, the system adapts.
Over time it learns.
That’s part of what makes the modern world so difficult to understand from inside it. Up close we see politics, markets, and crises. From farther away those events begin to resemble the behavior of a complex system—something with circulation, feedback loops, and a kind of nervous system made of communication networks.
And now we’re beginning to build machines capable of observing that system as well.
A few days ago I asked one of the newer AI systems a simple question.
“Why do global supply chains tend to stabilize economies?”
The answer was thoughtful and well organized. It talked about diversification, redundancy, and the way trade spreads risk across many regions.
But the interesting part wasn’t the answer.
It was how the machine managed to produce one at all.