Psst...there’s an audio version of this story. Upgrade to listen.

Urbit and the Not-So-Dark Future of the Internet

A dark genius is remaking the internet to revolve around the user

Image: Tlon, urbit.org

The internet is ostensibly a happiness engine. Studies prove that connection makes humans happy, perhaps happier than anything else, and the internet, at its core, is a connection machine. Yet studies also show that the more time we spend online, the more depressed we are. Surely, if the internet connected us the way we connect in real life, it wouldn’t make us so unhappy.

Solving this problem is not the target of many new apps; tech founders typically aren’t interested in acknowledging that most of what they produce doesn’t actually improve anyone’s lives. A solution will thus have to come from an outsider. Someone cynical enough to see the deep, fundamental flaws that make being online depressing and unreasonable enough to believe those flaws can be fixed.

Enter Curtis Yarvin and Galen Wolfe-Pauly, co-founders of a project called Urbit, which aims to fundamentally overhaul the internet by making it more like real life. If Urbit has its way, you’ll own not just all of your data, but all the data you interact with. For example, if you read a newspaper, you’ll keep that unchanging copy of the newspaper, just like you would in real life. If you have a photograph, it’ll belong only to you, not Mark Zuckerberg, the way a physical photograph would. Current dominant ecosystems like Facebook and Twitter will be reduced to something closer to photo carousels. At the same time, you won’t want to send hate emails from a fake Gmail account, because your nefarious actions will be much harder to separate from your real identity. Like in real life, your reputation will follow you everywhere.

Urbit will achieve this by presenting your whole digital life as a single web service. Every user gets a personal server that defines their entire online presence. These online selves will be indivisible from our real selves, and you will be able retrieve and distribute data on your own terms, just as you do every day in the real world.

Urbit represents what could be considered a Copernican Revolution of the internet.

Copernicus discovered that the Earth orbits the sun, not vice versa, which turned everything we thought we knew about physical space on its head. In the late 18th century, the Copernican Revolution in philosophy occurred when Immanuel Kant established that human perception does not revolve around objects, but that objects revolve around human perception. A chair looks like a chair because of how our brains filter information, not because of how a chair looks “in itself.”

Kant’s discovery revolutionized thought and, in a way, opened the door for computers, which prove daily that our perception can be easily manipulated by replicas of the natural world. Urbit may be a Copernican Revolution for the internet because, like Copernicus and Kant, it reverses the way we think about subjects and objects. It’s not us, the users, revolving around online ecosystems, but those ecosystems revolving around us.

This may sound like an obscure Elon Musk pet project, but it’s legit. Urbit’s primary goal, decentralization of the internet, will be one of the next big conversations, and several conferences and products have already popped up to address it. As the most ambitious of these decentralization projects, Urbit is backed by Founders Fund, Peter Thiel’s seed fund, as well as Tim Draper, Jaan Tallinn, and Andreessen Horowitz. A recent public crowd sale of Urbit address space sold out in four hours and raised more than $200,000. The disruption of all disruptions may actually be on the horizon.

Obligatory Yarvin Controversy Paragraph

Before we get into the actual important stuff — how Urbit works and why — we must unfortunately placate those who would reduce everything they don’t understand to political controversy. A lot of the buzz around Urbit comes from the infamous reputation of its aforementioned founder, Curtis Yarvin. The computer scientist, who made his name by developing very early smartphone software, rose to fame writing radical political philosophy under the pseudonym Mencius Moldbug. His most famous work is An Open Letter to Open-Minded Progressives, in which he, a lapsed San Francisco liberal, lays out an expansive critique of contemporary liberalism using a mind-bogglingly dense array of facts and references, from obscure 18th-century Naples to rapper 50 Cent. This work is credited with sparking the neoreactionary and Dark Enlightenment movements, which are in turn credited for sparking the “alt-right” anti-globalist movement.

Because of this, Yarvin’s invitation to speak about Urbit at the 2015 Strange Loop programming conference in St. Louis, Missouri, was rescinded after complaints from attendees. In 2016, five speakers and several sponsors withdrew from LambdaConf due to Yarvin’s inclusion.

This is almost entirely irrelevant to Urbit other than to point out one interesting irony. In his writing, Yarvin (very, very theoretically) argues that democracy should be replaced by a sort of neo-monarchy. With Urbit, he’s doing the precise opposite — liberating the individual in an internet dominated by feudalism.

Why the Internet Needs an Overhaul

Yarvin graduated from high school in 1988 and, like many computer geeks of his generation, yearns for the internet’s early days, when Usenet was a paradise and everyone got along. The internet was for smart people who wanted to change the world, and a strong code of honor reigned.

“The internet that I remember was a closed, very elite community. It was like the Burning Man constituency. You put the 100,000 best-educated people in the world in one spot, and they build something cool,” says Yarvin. “But when it doesn’t scale for ordinary human beings, you have a problem.”

According to Yarvin and Wolfe-Pauly, the problem with the contemporary web is that it wasn’t built to accommodate everyone in the world. Unix, the operating system upon which the internet is rooted, is essentially an industrial tool created by telephone companies in the 1970’s to operate mini-computers, and it hasn’t changed much since. Networks and programs were built on top of it without any idea that they would one day accommodate billions.

“There was no plan. No one knew the whole world would be relying on it,” Galen Wolfe-Pauly told me during a Skype interview with the duo that lasted close to three hours. Where Yarvin is your traditional techie, Wolfe-Pauly is young and fit with a shaved head. He plays the role of rational man while Yarvin plays the visionary. Both are extraordinarily passionate about their mission.

“Because of its primitive infrastructure, everything on the web was created as a sort of lightweight toy,” says Yarvin, who often refers to the major internet ecosystems as overgrown toys.

“Facebook is more or less a toy that somehow grew into a $500 billion company, serving 2 billion people off what’s essentially one giant server.”

This is evidence, Yarvin believes, that “we’re not living in an adult digital universe. More like a funhouse environment.”

Yarvin, like many computer scientists, realized early on that the shoddy, unplanned nature of the internet’s core systems would eventually cause big problems. After dropping out of a Berkeley PhD program in 1994, he began to develop the early ideas for Urbit. Yarvin thought that Unix and HTTP (the other vital inner layer essential to the modern internet) would eventually need to be paved over like the rough-hewn roads that they are.

“What I want to do is essentially pave over this industrial infrastructure,” says Yarvin. “The layer is going to be totally sealed off. Then we can basically design our software stack like we’re building on bare metal. This is the answer to the question of how we compute in the 21st century.”

But what is it exactly about Unix and HTTP that Yarvin finds so unsatisfactory? In answering this question, the Urbit team identified four problems their product is designed to address.

1. Namespace Mutability

One of the problems with the current internet is that it’s frustratingly impermanent. This is because its original infrastructure was built as a mutable namespace. A namespace is simply a network that maps names to identities, like a yearbook. An immutable namespace is one where neither the names nor the identities change. A paper yearbook is an immutable namespace because neither the names nor the photographs can change. Another example is a dictionary, which is an immutable namespace of the English language. Stable systems, including almost all that we deal with in the real world, are generally immutable namespaces. That’s because the human brain handles permanence much better than it does constant change.

The internet, however, is a mutable namespace. When you type in a URL and press return, the underlying webpage could be entirely different than it was five minutes or even five seconds ago. Users have no guarantee that important, world-changing headlines in major newspapers won’t be changed on the publisher’s whim. The New York Times was (inaccurately) accused of doing exactly this back in March. The Times has, however, “stealth-edited” articles in ways that effected the tone and substance of the article without noting such edits with a correction. In fact, when I was writing for LA Weekly, “stealth-edits” were the norm. After receiving complaints from subjects, I would, with the approval of my editors, frequently go into the publishing software and change the text after publication. Rarely if ever was a corresponding correction published.

“It’s as if the world is changing around you,” says Wolfe-Pauly. “You trust your memory. That’s how we build our intuition. And we can’t do that on the internet as is.”

Mutability is found in many forms throughout our online experiences, from broken links to deleted comments to tweaked photographs. It’s something we take for granted, but it turns out that it doesn’t need to be that way. The internet’s builders simply never considered how their decision to make its namespaces mutable would aff later generations if their creation spread to the entire world.

2. Feudal Ecosystems

If someone were to come into your house, take photos out of a box from under your bed, and sell them to an advertising agency, you would call the police. Yet you willingly accept it when Facebook does the same — not just with your photos, but with all of your personal data.

“You really can’t make a rule that says Mark Zuckerberg can’t have sex with your data, because the whole reason you’re there is that Mark Zuckerberg has sex with your data,” jokes Yarvin. He has an expansive sense of humor, which in print can read far less jovial than it’s delivered in person. “If I’m a genius creating content, I’m essentially slaving on the content plantation of someone else. And the thing is, the content plantation owner has no idea what he’s got.”

In exchange for a place to store and share your information, you provide content and engagement, which seems like a fair-enough trade. However, the problem is that the plantation owners currently have no restrictions on how they can treat you or your data. If they want to, they can delete your entire account for no reason at all. You can lose followings you’ve been building for years without any due process whatsoever.

Image: Tlon, urbit.org

“With the current system, you’re really seeing the virtues and vices of dictatorship. Facebook is very well operated and efficient. Imagine the U.S. government trying to run it. They couldn’t do it. It’s a little harder than running a health care website!” says Yarvin. “But also, you look at Twitter. I’ve heard it described it as a sewer, and now it’s a sewer with unexplained disappearances.” What Yarvin means is that, in a series of feudal dictatorships that comprise the contemporary internet, the monarchs are so powerful that they’re free to eliminate anyone they consider a threat without consequence.

Urbit wants to put an end to this Copernicus style by reversing how our relationship with these ecosystems works.

“If you think about where cultural knowledge lies, you have to ask, ‘Who owns culture?’ We’re trying to make it so that we all own culture,” says Wolfe-Pauly. He believes that if Twitter bans you for saying the wrong thing, you shouldn’t lose a single follower or a single tweet. You should be able to simply find another content carousel to organize your data.

3. Data Balkanization

Since these feudal ecosystems want you to spend as much time laboring on their, and only their, content fields, they set up barriers that make it difficult to retrieve your data. When you try to work around these barriers, they get angry. For example, when a startup called Power Ventures tried to make an integrated Facebook/Twitter news feed, Facebook sued the company into oblivion.

One particularly annoying manifestation of this is password fatigue. As more of our lives move online, it’s increasingly cumbersome to keep track of passwords. Password fatigue is like having a different key to every door to every room in your house, or, as one article put it, 295 different driver’s licenses.

For Yarvin, data balkanization is strong evidence that the current internet was never meant to be integral to our lives. You should be able to shop, bank, plan, and share without having to enter and exit so many different UIs.

“We really thought all these overgrown toys were the way the internet should be. All these companies’ market caps are based on the idea that this is the permanent solution, that Facebook will be there for your great-grandchildren. I’m not laughing at the work that’s gone into the Facebooks and Googles of the world, but I see no vision of the future in them. I just see a vision of how to take advantage of the present.”

4. Trollability

Part of the reason the internet sucks is that people can say whatever they want without fear of consequences. This is a function of mutability in combination with the ease of maintaining anonymity.

Computers are identified through IP addresses, but these tell you nothing about the identity of the user. They’re also easy to cloak. In real life, we are bound to our bodies and our reputations, both of which we cannot escape no matter how hard we try. But online, we are invisible and anonymous from the start. “One of the central mistakes was not to allocate IPs by person,” says Yarvin. “Instead, they’re allocated to ISPs by need, then distributed as the ISPs see fit.”

The social layer of the internet is similarly susceptible to trolls. Most free ecosystems like Google, Facebook, and Twitter do very little to ensure that the user is who they say they are, let alone a human being. The internet allows us to act like children, where nothing we say or do sticks, and, sure enough, that’s exactly what we do.

How Urbit Will Fix These Problems

Urbit plans on solving these four problems by paving over the bumpy old infrastructure of the internet and building a new universe on top of it.

“One way to look at Urbit is like pavement. The old ground, essentially Unix, is just rocks and dirt,” explains Yarvin. “It’s incredibly difficult to travel over and build things on. Urbit is building a layer of smooth, simple asphalt on top of the jagged surface of Unix that will, first and foremost, make navigating much simpler and easier.”

Then, on that smooth and immutable new surface, Urbit will create a new internet universe where real humans are each bound to their own individual servers. Thus, instead of your IP address revolving around the massive single server that is Facebook, Facebook will revolve around a small single server that is you. All data on Facebook — yours and anyone else’s (who’s given you access) — will be freely retrievable and shareable anywhere. You’ll be able to see all your social feeds in one place and share content across all platforms at once. Instead of a feudal ecosystem owning your data, you maintain it all on your Urbit and use built-in software to scrape existing ecosystems to retrieve data and display it how you want. Will this eventually make the ecosystems themselves obsolete? Almost definitely. But in the meantime, Urbit believes it can pave over them in a way that puts you in control.

Since Urbit’s namespace will be immutable, the “names” in its galaxies will refer only to static identities. Imagine that you want to read today’s New York Times. You’ll simply tell your Urbit to retrieve it for you, and it will deliver an immutable copy of the New York Times for that day and time. If the New York Times goes back and changes anything in its database, that change will be both obvious and irrelevant, because you will own the copy you retrieved, just like you would own a paper newspaper in real life (that is, “New York Times, April 5, 2017”). Thus the mutability problem is solved. No more broken links, no more deleted comments, no more edited headlines. You own an immutable version of any data you choose to retrieve via your Urbit.

Another thing about Urbits is that they cost money, somewhere around $5 to $20 a month for one of 4 billion. This is how Yarvin and Wolfe-Pauly seek to raise money for the project and how they will enforce the integrity of the Urbit universe.

“It’s like getting your own phone number,” says Yarvin. Everything you do with this “phone number” is recorded by the other Urbits you interact with. “If you start sending spam, that will follow your Urbit everywhere it goes. And you’re out 20 bucks, because you have to buy a new Urbit.” This goes a long way toward solving the problem of trollability.

That said, if you buy two or more Urbits, you’ll still be able to stay anonymous on one or more of them. However, the more you troll on an anonymous Urbit, the fewer places it will be able to go, because it will appear on troll databases created by other users designed to exclude troll Urbits from interacting with their Urbits. Remember, your reputation follows you everywhere. That will get expensive very quickly as you wear out Urbit after Urbit.

What Will It Look Like?

The closest current reflection of what the Urbit UI will look like is the Chinese app WeChat. On WeChat, you are the master of your online destiny. Instead of doing the work to integrate with existing ecosystems — banks, social apps, maps — WeChat puts the onus on the ecosystems themselves to format their data in a way that integrates seamlessly with WeChat’s UI. The result is that you can do literally everything in one place. You can do the Chinese equivalent of a tweet, order food, pay a friend, invest in a stock, send a text message, share a photo across multiple social accounts, and even pay for a meal in a restaurant, all in one UI and without entering password after password.

So why not just make a U.S. version of WeChat? Because the refusal of our large ecosystems to cooperate with each other makes this kind of centralized UI impossible. Furthermore, WeChat is just one very large and effective feudal ecosystem. It’s better at communicating for you, but WeChat still owns all your data and can do whatever it wants with it. It decentralizes power from existing ecosystems, only to deliver it to an even more centralized ecosystem. Urbit, on the other hand, decentralizes the power of ecosystems by centralizing power with individual users.

But if there’s no centralized governing body to control things, how will Urbit’s universe be organized? How will it enforce its rules without becoming like the tyrants it’s trying to disrupt?

The Urbit Universe

The Urbit universe will be divided into roughly 255 (2⁸) galaxies, 65,000 (2¹⁶) stars, and 4 billion (2³²) planets, with 2³² moons for each planet. Each planet will be an individual human, and each moon will be a device tied to that individual. Stars are Urbits with the ability to distribute planet Urbits to individuals, and galaxies are Urbits with the ability to distribute star Urbits. The system will be self-authenticating, with galaxies authenticating stars and stars authenticating planets. Stars and galaxies are just individuals who have paid for special Urbits that possess the power to authenticate other Urbits.

“Parent stars help deliver your information, but this has to stay compatible with digital freedom, or else we return to the digital plantation model,” says Yarvin. To ensure that stars and galaxies don’t simply become feudal ecosystems in their own right, he’s developed a system that severely limits their power. “If I’m a star and I sell you a planet, I can’t take it back. Second, if you’re a planet and your star is denying you service for whatever reason, you are free to move to another star—one of 65,000.”

While the outer limits are still largely theoretical, Urbit is very real and is technically already live. Planets and stars have been purchased: Besides last June’s sale of $200,000 in Urbits, Yarvin sold off about $2,000 worth back in 2013, because, as he says, “I was completely broke.” Urbit.org currently operates on an Urbit, and simple self-owned web publishing is already available.

Of course, other systems, like Git, IPFS, Ethereum, and even Bitcoin, are trying to make the web more decentralized, stable, mutable, and navigable in similar ways to Urbit. However, Urbit is the most ambitious and wide-sweeping of them. It is a total system overhaul, a potential Copernican Revolution.

“The distinguishing characteristic of Urbit is that it’s not a fix. It’s a whole new car,” says Yarvin. “None of these other things solve the whole problem of decentralization at once.”

Urbit recognizes that overambitious projects to fundamentally change the entire internet are not new. Yarvin and Wolfe-Pauly realize that they’re still eons away from even the early stages of rolling out the above-described product. So, for the time being, they’re focused on building a small niche.

“Pretty much every project like Urbit tries to accomplish the goal of ‘reinventing the internet’ in the traditional 18-month startup timeline,” says Wolfe-Pauly. “Not only has Curtis been working on Urbit for a really long time, but we also aren’t moving too quickly. Before we can compete with WeChat, we need to host a community of crazy nerds who believe in a project like this: exciting for the kinds of people who think Mastodon is cool or Ethereum is cool. The near-term future for Urbit is more like Usenet than WeChat.”

For the rest of us, however, the idea of Urbit represents a revolutionary and potentially Copernican break from the norm of a quite depressing internet. Unlike the original internet, Urbit is purpose-built to hold up for centuries under the weight of a massive population. In that way, it’s a bit like the constitution of a brave new country, one that’s learned from history what not to do. It’s perhaps the first modern, intentional attempt to structure a whole-internet government centered around the user.

“This is cultural infrastructure,” says Wolfe-Pauly. “We need this for society.”

Using his critic’s eye and trademark humor, Yarvin contrasts Urbit’s vision with that of the master of all feudal ecosystems.

“Google’s mission statement is ‘organize the world’s information.’ What that assumes is that the world’s information is a shitshow, a pile of crap, and we need to pay a bunch of people $300,000 a year to go in and sort it out. Our idea is that the world should organize its own information.”

Is that how a Silicon Valley entrepreneur usually talks? Definitely not. But perhaps it’s about time they started.