For almost all of human history, nobody did everything themselves. Even in a tribe of twenty, someone hunts, someone sews, someone builds. Specialization is older than money, older than cities, older than writing. We have always split the work and traded the results, and that trade is the part of the story that actually matters.

Picture society as a pyramid. Energy sits at the base, holding up everything above it. On top of energy you get the four things that turn raw power into a functioning life: manufacturing, farming, construction, and transportation. For most of history we did all of it by hand. We forged by hand, hunted and gathered by hand, built by hand, and moved goods on foot. Trade is what took that output and multiplied its value.

It started inside a single tribe, then ran between neighboring ones, then stretched across continents. The Silk Road is the famous version. Those routes were already ancient, but the Mongols took control of most of Eurasia in the thirteenth century and made them safer and busier than ever. The result was an unprecedented flow of goods and ideas across the known world.

From there it is one long process of friction leaving the system. The shipping container collapsed the cost of moving anything across an ocean. The bill of lading turned a shipment into a piece of paper you could trust. Money did the same thing for value itself: we went from trading berries for meat, to silver that held its value but was a pain to carry, to paper that was lighter, to digital that is weightless and instant. Each step took friction out of the exchange.

That is still the direction of travel. Make the supply side more efficient, then make the matching more efficient. Amazon is supply-side efficiency at industrial scale, at least in the first world. Meta built the other half, the matching layer, where people post what they are offering, advertise it, and get connected to the right buyer. Underneath it sits an enormous record of what specific people actually want.

AI is the next layer on top of all of it. In plain language, through a chat box, you get matched to what you want faster than any earlier tool could manage, whether that is a business looking for a partner or a consumer looking for a product. It is the same mechanism we have been refining for millennia, now with less friction in it.

The next step follows directly from the pattern. Before long you will close your laptop and your AI will already know, in general terms, what you want. It will go out, talk to other AIs, find the matches, and bring them back, so most of the proactive searching goes away.

I feel the gap every day. I run a logistics consultancy, and I still spend real time posting, cold emailing, and grinding on sales. There are probably thousands of companies out there with a need that fits what I do almost perfectly, and even now, with search and social media and every tool built to help people find each other, finding them is slow. That is the gap AI closes, especially once it goes agentic and starts doing the work without me looking over his shoulder.

There are plenty of uses for AI that have nothing to do with this. I am writing about one of them: supply meeting demand with less and less friction. It is the oldest job in the economy, and the one AI is set up to do better than anything before it.