On HBO's Silicon Valley, startups promise to "change the world" by tackling silly, often nonexistent problems. But this season, the show's characters are tackling a project that really could make a difference. In their latest pivot, Richard Hendricks and the Pied Piper gang are trying to create a new internet that cuts out intermediaries like Facebook, Google, and the fictional Hooli. Their idea: Use a peer-to-peer network built atop every smartphone on the planet, effectively rendering huge data centers full of servers unnecessary.
"If we could do it, we could build a completely decentralized version of our current internet," Hendricks says. "With no firewalls, no tolls, no government regulation, no spying. Information would be totally free in every sense of the word."
But wait: Isn't the internet already a decentralized network that no one owns? In theory, yes. In practice, a small number of enormous companies control or at least mediate much of the internet. Sure, anyone can publish whatever they want to the web. But without Facebook and Google, will anyone be able to find it? Amazon, meanwhile, controls not just the web's biggest online store but a cloud computing service so large and important that when part of it went offline briefly earlier this year, the internet itself seemed to go down. Similarly, when hackers attacked the lesser-known company Dyn—now owned by tech giant Oracle—last year, large swaths of the internet came crashing down with it. Meanwhile, a handful of telecommunications giants, including Comcast, Charter, and Verizon, control the market for internet access and have the technical capability to block you from accessing particular sites or apps. In some countries, a single state-owned telco controls internet access completely.
Given those very non-utopian realities, people in the real world are also hard at work trying to rebuild the internet in a way that comes closer to the decentralized ideal. They're still pretty far from Richard's utopian vision, but it's already possible to do some of what he describes. Still, it's not enough to just cut out today's internet power players. You also need to build a new internet that people will actually want to use.
On the show, Richard's plan stems from the realization that just about everyone carries around a smartphone with hundreds of times more computing power than the machines that sent humans to the moon. What's more, those phones are just sitting in people's pockets doing nothing for most of the day. Richard proposes to use his fictional compression technology—his big innovation from season one—to free up extra space on people's phones. In exchange for using the app, users would agree to share some of the space they free up with Pied Piper, who will then resell it to companies for far less than they currently pay giants like Amazon.
The closest thing to what's described on Silicon Valley might be Storj, a decentralized cloud storage company. Much like Pied Piper, Storj has built a network of people who sell their unused storage capacity. If you want to buy space on the Storj network, you upload your files and the company splits them up into smaller pieces, encrypts them so that no one but you can read your data, and then distributes those pieces across its network.
"You control your own encryption keys so we have no access to the data," says cofounder John Quinn. "We have no knowledge of what is being stored."
Also like Pied Piper, Storj bills itself as safer than traditional storage systems, because your files will reside on multiple computers throughout the world. Quinn says that in order to lose a file, 21 out of 40 of the computers hosting it would have to go offline.
Storj proves that the Silicon Valley's basic idea is feasible. But unlike Pied Piper, Storj doesn't rely on smartphones. "Phones don't have much storage and the network capability isn't great, so the show's idea is a little fanciful," says Quinn. Someday, 5G wireless networks might make phones a more viable part of the Storj network. If Richard's compression algorithm was real, those smaller files will help too. But for now, the Storj network relies primarily on servers, laptops, and desktop computers. The reality is less grand than the HBO fantasy.
As interesting as Storj is, it's not quite what Richard actually described in his pitch. Storj is a storage service, not a whole new internet. A more ambitious project called IPFS (short for "Interplanetary File System") is probably a bit closer to Richard's grand vision of a censorship-resistant internet with privacy features built right in.
The idea behind IPFS is to have web browsers store copies of the pages they visit and then do double-duty as web servers. That way, if the original server disappears, the people who visited the page can still share it with the world. Publishers get improved resilience, and readers get to help support the content they care about. With encryption a part of the protocol, criminals and spies can't in theory see what you're looking at. Eventually, the IPFS team and a gaggle of other groups hope to make it possible to build interactive apps along the lines of Facebook that don't require any centralized servers to run.
But the idea of building a censorship-proof internet by backing up copies throughout the internet isn't without its potential problems. Sometimes publishers want to remove old content. IPFS creator Juan Benet told us last year that the project is trying to work out ways to let publishers "recall" pages that are being shared. But that idea is also fraught. What's to stop a government censor from using the recall feature? What happens if someone creates a version that ignores recalls?
Then there are moral and legal risks. Tools like Storj and the venerable peer-to-peer sharing system Freenet make it impossible to know just what content you're storing for other people, which means you could be playing host to, say, child pornography. Quinn says that the Storj team is currently working on ways to block known problem users. But it won't be able to completely guarantee that none of its hosts will end up storing illegal content.
IPFS gets around this largely by letting people decide which of the content they've visited they actually want to share. But this means that less popular content, even if it's perfectly legal and ethical, might end up disappearing if too few people share it. Benet and company are working on a system called Filecoin that, not unlike Storj, would compensate people for providing access.
Even overcoming these trade-offs inherent in decentralization, people may still not want to use these apps. Storj may be able to win over businesses by being cheaper, but even if it is more reliable, the idea of storing data on random machines scattered across the internet instead of in a traditional data center sounds risky compared to, say, the massively robust AWS, backed by Amazon's technical know-how and billions of dollars. Convincing people to use decentralized alternatives to Facebook and Twitter has proven to be a notoriously difficult problem. Getting people to use what amounts to a whole new version of the web could be even harder.
Even if IPFS, Storj, or one of the countless other decentralized platforms out there do win people over, they're still technically riding atop the existing internet infrastructure controlled by a shrinking number of telcos. Silicon Valley hasn't addressed this problem yet. But what if you could chain the smart phones and laptops of the world together using WiFi and Bluetooth to create a wireless network that was free and open to everyone, with no need for Big Telecom?
Australian computer scientist Paul Gardner-Stephen tried to do something like that after the Haiti earthquake in 2010. "Mobile phones have the capability to run autonomous networks, it's just that no one had implemented it," he says. Gardner-Stephen helped build Serval, a decentralized messaging app that can spread texts in a peer-to-peer fashion without the need for a traditional telco carrier. But he quickly realized, as the Pied Piper team likely will, that trying to turn people's mobile phones into servers drains their batteries too quickly to be practical. Today, the Serval team relies on solar powered base stations to relay messages.
Serval and similar apps like Firechat aren't meant to replace the internet, just provide communications during disasters or in remote locations. But the idea of creating decentralized wireless networks—mesh networks—still has merit. One such network, Wlan Slovenija, for example, now covers all of Slovenia and is spreading to neighboring countries. But these mesh networks are still a long way from replacing telcos—especially in the US. Even as wireless base stations improve, they can't quite yet compete with the fiber-optic cables that link the nation's telco infrastructure on speed and reliability, and some community networks, such as Guifi in Spain, are bolstering their wireless connections with fiber.
Even then, given a choice, would people really pick a decentralized option over the status quo? Customer service at big broadband companies may be bad to nonexistent, but you can still call someone. For those who would nevertheless prefer to wrest control of the internet from large corporations, these new alternatives will need to be better and faster than the services they hope to displace. Simply being decentralized isn't enough.
It wasn't so long ago that people questioned whether anyone would ever take to the internet at all. As the season finale of Silicon Valley approaches, Pied Piper will find out whether its version of a new internet works and whether there's a demand for it. They just have to build it and see if anyone comes—just like in the real world.
