Last night on "60 Minutes," Amazon CEO Jeff Bezos showed Charlie Rose a video from the future.

A drone with propellers, nicknamed "octocopter," snatched a package off a conveyer belt in an Amazon warehouse, took off through the door, flew above a grassy field, and dropped it off in a family's backyard. It's a neat clip (watch it below), and within five years, Bezos said, Amazon PrimeAir won't just be YouTube video. It will be a way to deliver packages to families within 30 minutes of their purchase click.

The project is theoretical bordering on science fiction, and five years is an eternity in Internet and mobile technology (five years before Apple introduced the phone I use, iPhone 4S, there was no iPhone). But in a way, Amazon's drone dreams fit snuggly into the sweep of retail innovation, which I reviewed for my last Atlantic column, The Amazon Mystery.

The turbulent 150-year history of American retail is one of enormous companies desperately trying to anticipate the whims of mobile shoppers and mobile technology. Amazon's stated plan to ask robots to deliver packages in cities is a perfect extension of the last century's lessons in how retail companies grow, thrive, and die.

Planes, Trains, and Octocopters

In 1890, the frontier was "closed," and for the first time it could be said that the union was united—by rail and telegraph. Although the population was still two-thirds rural, mass retailers like Sears and Montgomery Ward could respond to catalog orders, load merchandise onto outgoing trains, and accomplish something unprecedented. They brought the stores to the people.

In the early-to-mid 20th century, however, a "mobile" technology even more advanced than trains—the automobile—did something even better: It brought the people to the stores. Sears adapted to this transportation shift brilliantly. As families streamed off farms into cities, Sears followed, building 300 national stores in four years. With revenue approaching 1 percent of the entire U.S. economy, Sears was America's perfect retail company.

Until it wasn't. In the 1980s, the slow demise of manufacturing weakened Sears' core demographic. Meanwhile, new competitors were lapping Sears in the race for the best information technology, both down-market, like Walmart, and up-market, like The Gap. From their brief history of the rise and fall of Sears, Daniel Raff and Peter Temin write:

For Sears to have competed directly with Wal-Mart, it would have had to rethink from the ground up how goods passed from manufacturers to consumers. Sears managers told themselves that Wal-Mart and other low-price firms succeeded because they sold cheap goods, representing a move down-market that Sears would not follow. The focus on the goods sold obscured the innovations in the way Wal-Mart and other firms organized like it handled the products. The goods were cheap partly because Wal-Mart's costs were low.

In short, Sears lost on infrastructure.

When people think of retail companies, what comes to mind is everything we can see inside the store: the merchandise, the layout, the clientele. But more important is everything you can't see inside the store: the supply chain, the warehouse management, the software that relays sales data to the people making decisions about the next batch of orders. In the short run, shopping is about seizing on trends and stocking the most popular merchandise. In the longer run, the best retail companies turn out to be the best infrastructure companies.

And drones are an answer perhaps the most contentious retail/infrastructure question facing Amazon and its competitors: Convenience and delivery time.

Drone Dreams

"Amazon is an e-commerce platform and a logistics machine," the analyst Benedict Evans emailed me for my research on the Atlantic column. "This is what Sears did with the mail and railways. Amazon is doing it with the web and now mobile."

Today, Amazon is the world's most successful online retail company because it has the world's best offline, real-world infrastructure for rapidly getting stuff to people's doorsteps. But the competition is steep: Walmart, eBay, and a batch of startups are breathing down its neck, and that's just in the U.S. The company faces even stiffer competition in Europe and China.

Just as small-town stores couldn't have predicted how the continental railroad would bypass them, and Montgomery Ward couldn't have predicted how cars would diminish the mail-order business, and Sears didn't anticipate the technologies that would catapult Wal-Mart, it's hard to predict what development will mark the next inflection point in online retail.

Bezos' plan to fill the skies with drones might be a PR stunt. But it's also a vision of distribution that fits perfectly with their current strategy to cut order times for downtown homes. The company, which used to buy warehouses in cheap states to avoid taxes, is now building new centers near cities to win the delivery-time war for affluent urban customers. Drones would give Amazon an edge in speed and cost for items up to five pounds, which make up 86 percent of the items it delivers.

The potential for drone delivery is massive, but also, I think, inchoate. What sort of purchases are created or destroyed by having 30 minutes between you and almost every item in the Amazon catalog? How many drives to the store are eliminated? How many shopping stores and sites are obliterated by the convenience, and how much could Amazon earn charging tens of millions of households for such the service? And then, the consumer surplus: How much time and energy is saved by having robots fly the Everything Store to our doorstep? (Not to mention, how many jobs are eaten by the drones?) 

It's not worth going too deep on regulatory and social hurdles for a technology that might or might not exist in half a decade.  Rather, Amazon's drone dreams are another chapter in two old stories: Thinking long-term about infrastructure is what makes and destroys retail giants; and Bezos' perspective is nothing if not long-term.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.