Deeply strange reports have been emerging from the Las Vegas headquarters of Zappos, until recently the world’s happiest shoe store. This spring, by order of the CEO, Tony Hsieh, the company abolished managers, eliminated job titles, denounced its own organizational hierarchy, and vested all authority in a 10,000-word constitution that spells out a radical new system of self-governance. Holacracy, it’s called, and it makes all previous moves toward “employee empowerment” look like the mild concessions of an 18th-century monarch. Freed from direct supervision, employees are expected to join various impermanent democratic assemblies called “circles” (headed, but not run, by a “lead link”), in which they will essentially propose their own job descriptions, ratify the “roles” of others, and decide what projects the group should undertake.

The constitution was written by Brian Robertson, a Philadelphia entrepreneur unaffiliated with Zappos who’d grown disaffected with standard management practices. Hsieh had heard him speak about holacracy at a 2012 conference, and was captivated. “We adopted it wholesale with zero changes,” says John Bunch, who used to be a technical adviser at Zappos but is now title-less. (He is, nevertheless, heading the implementation of holacracy.)

The media has so far reacted with a mix of understandable skepticism and outright derision. “Mr. Hsieh’s hot hand appears to be at risk of going cold,” The New York Times reported in July. Employees at Zappos, the paper said, have met the development “with everything from cautious embrace to outright revulsion.” The tech site Pando was less measured: “Holacracy of Dunces,” a headline snorted that same month.

Holacracy’s implementation at Zappos, still in process, has undoubtedly caused problems (more on those later). But such reports risk missing the larger picture. However fraught it may be, Zappos’s experiment with holacracy is just the latest sign that information technology is allowing the emergence of a new form of organization.

For years, pockets of the U.S. military have been slowly taking decisions out of the hands of high-ranking commanders and entrusting them to teams of soldiers, who (armed with the concept of “commander’s intent”) are told what problems to solve—but not how to solve them. “The organization as a rigidly reductionist mechanical beast is an endangered species,” General Stanley McChrystal writes in his new book, Team of Teams. “The traditional heroic leader may not be far behind.” At the video-game maker Valve, new employees are told not to expect instructions, because even the managing director “isn’t your manager,” says the employee handbook. “You have the power to green-light projects. You have the power to ship products.” And so they do.

What’s enabling this shift, argues Thomas Malone, a professor at MIT’s Sloan School of Management, is simple: falling information costs. In his 2004 book, The Future of Work, Malone broke the history of organizations into three stages. In stage one, information is expensive to convey, so most decisions are made face-to-face in necessarily small firms. As communication costs begin to fall, Malone explained to me, we reach stage two, in which “it becomes economically feasible to send information to a single, central place for decisions to be made.” That is, the large, centralized hierarchy becomes possible. Then comes stage three: “As communication costs continue to fall, there comes a time when it’s economically feasible to bring information to all points, so in some sense, everyone can know everything.” In this third stage, the benefits of bigness can persist, but its traditional handmaiden, hierarchy, doesn’t have to. (Indeed, when the volume of information grows large enough, trying to direct its flow upward for evaluation can slow everything down.)

Malone points to the history of government. We lived most of our existence in small, fairly egalitarian bands of hunter-gatherers, until the advent of written language, which arose in tandem with record-keeping, taxation, and the founding of the first kingdoms, around 3000 B.C. Large-scale democracy did not appear until an equally momentous information technology, the printing press, enabled the “reading revolution” of the 18th century.

In business, the transition from bands to kingdoms came courtesy of the telegraph in the mid-19th century. And what it begat—the large, managerial hierarchy; the salaried manager; the modern definition of a job—has endured, with occasional pushes to prune and flatten management, until now, a good two decades after the Internet radically altered the volume, velocity, and direction of information flows. (Doubters, check your inbox.)

Why the lag? A look back at the mid-19th-century transition suggests an answer. After a new information technology appears, the organizational innovation that it allows doesn’t knock at the front door and present itself. People have to invent it. The ones who did last time around aren’t as celebrated as Morse, or before him Gutenberg. But what they arrived at looks obvious only in retrospect.

In 1854, the board of the New York and Erie Railroad promoted Daniel McCallum, a Scottish-born bridge builder, to general superintendent, hoping he could address the railroad’s mysterious ills. As the railroad grew larger, it was growing not more efficient per mile, as Adam Smith would have predicted, but less so. Had the company hit some sort of natural size limit? No, McCallum concluded. What was breaking down was its stage-one organizational system.

At a small railroad—and all railroads up to then had been small—the superintendent could give every aspect of its business his personal attention, McCallum wrote. “Each employee is familiarly known to him, and all questions in relation to its business are at once presented and acted upon.” But as The Atlantic marveled in an 1858 article featuring McCallum’s ideas, the New York and Erie had to coordinate the movements of some 200 locomotives, 3,000 cars, and 4,700 employees dispersed over hundreds of miles. What was needed, McCallum wrote, was “a system perfect in its details, properly adapted and vigilantly enforced.”

The system he spelled out—in great detail—delineated who had the authority to decide what, and even what words to use in communicating that decision back to their superiors on the telegraph. (“All subordinates should be accountable to, and be directed by their immediate superiors only.”) In the elaborate diagram that McCallum created—widely considered the first modern organizational chart—17 spokes, each representing a different division, radiate from a central hub that is the office of McCallum himself. Thanks to a system of hourly reports, The Atlantic wrote,

the General Superintendent at his office can at any moment tell within a mile where each car or engine is, what it is doing, the contents of the car, the consignor and consignee, the time at which it arrives and leaves each station, (the actual time, not the time when it should arrive,) and is thus able to correct all errors almost at the moment of commission … The great regulator … is the electric telegraph, which connects all parts of the road, and enables one person to keep, as it were, his eye on the whole road at once.

The most striking aspect of McCallum’s system comes near the top of his report. For all the rigidities, he admitted, the system was an improvisation—one man’s attempt to bring his organization into better alignment with a rapidly changing environment. It would “require amendment,” he wrote, “and a reasonable time to prove its worth.”

Which sounds a lot like the commentary of Brian Robertson, the entrepreneur who created holacracy. Like McCallum’s, his decision to put pen to paper was not utopian but pragmatic. The idea was born out of his own frustrations as the CEO of Ternary Software, which he founded in 2001. “When I tried to design organizations purely from my own intellect,” Robertson says, “I always got it wrong. I’d design solutions for problems that didn’t exist. Other parts were overdesigned. We’d end up solving problems that would never come up in practice, and then you’d miss the ones you actually needed to solve. You can’t know enough.”

He didn’t start with any radical notions of overturning hierarchy. He just wanted his company to function better. After exhausting the most-obvious solutions—hiring better managers, attempting to nail the perfect org chart—he finally concluded that the problem was the very concept of a rigid org chart. And so, as McCallum had a century and a half before, he wrote out a new set of rules delineating a new system.

Like McCallum’s, it is very, very complicated—too complex to describe fully here, and too complex for the liking of some at Zappos. (The constitution Robertson drafted runs more than twice the length of the Founding Fathers’.) But its basic principles go something like this: Instead of being assigned job descriptions from on high, employees take on mutable roles where they see a need. The hierarchical org chart is replaced by an ever-changing array of circles that can form, merge, or collapse in response to opportunities and threats in the marketplace. Reorganization, in effect, becomes a permanent way of life. The rigid forms invented in the steam age give way to something more organic—a jellyfish, perhaps.

It makes some sense on paper. But in practice? Holacracy’s first laboratory, Ternary, was a tiny company that didn’t survive the Great Recession. And at Zappos, given the choice of embracing holacracy or taking a buyout this past spring, 210 employees, or 14 percent of the company’s workforce, chose the latter. John Bunch notes that holacracy has created a whole set of new problems that have to be addressed. How to evaluate people’s performance is one. (Zappos is working on a system called “badging,” whereby employees earn badges for acquiring new skills.) How to compensate someone who splits her time among three different roles (say, customer care, event planning, and brainstorming a new Zappos product line) is another. The list goes on. How do you stop people from trying to exercise power they no longer have? And how does a system conceived to reduce information overload (“the goal,” says Bunch, “is to find the optimal circle structure where the need for communication between the circles is the least”) not add to it instead, in the form of endless meetings?

But just because solutions don’t yet exist doesn’t mean they won’t be found, says Robertson, who likens holacracy to a computer’s operating system and the solutions to apps. Remember when the App Store first opened? It didn’t have a lot on offer. Bunch told me that it’s too soon for Zappos to begin making amendments to holacracy; the company needs to finish the difficult process of understanding it first. “If someone comes to me super-frustrated, borderline angry,” he says, “that’s actually a really good sign to me, because it means they’re at least trying to understand it.”

The important point is this: To pronounce holacracy unworkable now would be akin to pronouncing democracy unworkable in the midst of the French Revolution. And should holacracy at Zappos fail, which it well may, neither should its principles be pronounced dead. The New York and Erie Railroad was in receivership within a few years of McCallum’s report—but the system he pioneered there became utterly commonplace by century’s end.

With that in mind, it’s not hard to imagine a future in which the only thing strange about what’s going on at Zappos is that it ever seemed strange at all.