As entities on the Web, we have devolved. Client-server has become calf-cow. The client—that's you—is the calf, and the Web site is the cow. What you get from the cow is milk and cookies. The milk is what you go to the site for. The cookies are what the site gives to you, mostly for its own business purposes, chief among which is tracking you like an animal. There are perhaps a billion or more server-cows now, each with its own "brand" (as marketers and cattle owners like to say).

This is not what the Net's founders had in mind. Nor was it what Tim Berners-Lee meant for his World Wide Web of hypertext documents to become. But it's what we've got, and it's getting worse.

In February 2011, Eben Moglen gave a landmark speech to the Internet Society titled "Freedom in the Cloud", in which he unpacked the problem. In the beginning, he said, the Internet was designed as "a network of peers without any intrinsic need for hierarchical or structural control, and assuming that every switch in the Net is an independent, free-standing entity whose volition is equivalent to the volition of the human beings who want to control it". Alas, "it never worked out that way". Specifically:

If you were an ordinary human, it was hard to perceive that the underlying architecture of the Net was meant to be peerage because the OS software with which you interacted very strongly instantiated the idea of the server and client architecture.

In fact, of course, if you think about it, it was even worse than that. The thing called "Windows" was a degenerate version of a thing called "X Windows". It, too, thought about the world in a server-client architecture, but what we would now think of as backwards. The server was the thing at the human being's end. That was the basic X Windows conception of the world. It served communications with human beings at the end points of the Net to processes located at arbitrary places near the center in the middle, or at the edge of the Net. It was the great idea of (Microsoft) Windows in an odd way to create a political archetype...which reduced the human being to the client and produced a big, centralized computer, which we might have called a server, which now provided things to the human being on take-it-or-leave-it terms.

True, but let's not forget Netscape's role. The HTTP cookie was created by Lou Montulli in 1994, when he was working at Netscape on the first e-commerce servers. The idea was innocent enough: to recall state and save work by remembering shared data, such as the contents of shopping carts. Today, cookies are used for many other things, including user tracking, mostly so advertising can be personalized. The less-polite word for this is spying. Eben singles out Facebook as an especially egregious offender and describes its offering this way: "I will give you free Web hosting and some PHP doodads and you get spying for free all the time".

This trade-off is the rule, not the exception. In the study that launched its "What They Know" series last year, the Wall Street Journal said all but one of the 50 most-popular Web sites installed tracking cookies in users' browsers. (The only exception, no surprise, was Wikipedia.) One, Dictionary.com, installed 234 tracking files, of which all but 11 were from tracking companies.

The end state of this system was forecast with metaphorical precision by The Matrix, back in 1999. From the standpoint of its business model, the advertising-supported commercial Web is a machine world in which users are nothing more than batteries to whom "experience" is "delivered" through pipes. In that world, advertising companies are the agents, and tracking cookies are like the electric scorpion that wiggles into Neo through his navel so he can be followed around.

Eben doesn't want us to go there. So he had a plan: FreedomBoxes. These are "cheap, small, low-power plug servers...the size of a cell-phone charger, running on a low-power chip". He sees these boxes starting cheap and dropping in price, eventually costing less than $30 apiece. That's roughly the price of cell-phone earbuds.

Thus, Eben and friends created the FreedomBox Foundation, set up a KickStarter project and quickly attracted enough donations to get rolling. (They asked for $60,000 and got $86,724.) Others jumped on board. Debian, for example, has its own FreedomBox project to support the foundation's efforts with working code.

Although Eben didn't aim his appeal to big companies, we can use their help, especially at the retail level. For FreedomBoxes to be popular, they'll need to show up where the populace shops. That means we should see FreedomBoxes on the shelves at Best Buy, Radio Shack and Walmart—and not just at, say, Amazon. Hey, companies don't need to be cows any more than humans need to be calves. When the world fills up with FreedomBoxes, customers can help companies evolve.

The economic case we need to make is that free customers are more valuable than captive ones. But, we can't make it if we're still on the ranch. We need to break out.