In defense of decentralization

Call it the Battle of the Princeton Tigers.

Arvind Narayanan is an associate professor of computer science at Princeton University, digital privacy expert, and overall smart dude. He’s researched and taught courses on cryptocurrency and blockchain since 2013, but has been cautioning the public against the “pitfalls of decentralization” for almost as long.

Joseph Lubin is the founder of ConsenSys (Decrypt’s source of funding), a graduate of Princeton, and an equally smart dude. He co-founded Ethereum in 2014 and has remained steadfast in his belief in a societal need for decentralization.

Let’s rumble.

On New Year’s Eve, Narayanan took to Twitter to pronounce that the “blockchain/decentralization story fell apart” in 2018. He then went on to deliver one of the more measured and salient critiques of the current state of blockchain in recent memory. In a tweetstorm that generated dark clouds over Crypto Twitter, Narayanan deconstructed what he views as “leaps in logic” regarding the perceived relationship between blockchain technology (that is, software) and the decentralization of society.

Narayanan’s argument boils down to the idea that centralization—contrary to popular belief—isn’t necessarily the result of government overreach or “monopolistic rent-seeking.” Rather, he said, it is a natural consequence of free-market interactions “due to economies of scale and other efficiencies.” To illustrate his point, he cited the parallels in the evolution of bitcoin mining and gold mining—from individuals panning for resources in shallow streams to giant, centralized operations that require significant capital to run.

But we can see other more glaring examples in our everyday life—take Amazon, for instance. It did Wal-Mart one better and, through a combination of frictionless, pioneering online retail innovations, coupled with massive, centralized warehouses and efficient supply chains, threatens to eat the world. (It is perhaps not coincidental that this too involves the story of a Princetonian.)

Technology itself isn’t enough to produce the sort of decentralization that most blockchain enthusiasts seek, says Naranyanan. Instead, he claims, culture precedes software—meaning that technology simply adheres to the organizational structures that most people in a society desire:

“If [centralization] were a conspiracy,” he said, “then blockchains might be a good answer.” But that isn’t the case. “What we’ve got is what people want,” Narayanan wrote to his followers, and any significant move towards decentralization will take decades of slowly and organically fostering “an ecosystem of services that build on each other and gradually improve in functionality and quality.”

Lubin, as you’d expect, sees things differently, but doesn’t disagree that these sorts of changes take time. “I think lots of us are saying that, too,” Lubin tells Decrypt. “We’ll make steady progress, and building layer upon layer, and it’s already succeeding in some interesting ways.” But as for the inevitability of centralization as ecosystems grow in scale—or even the idea that centralized structures are more socially desirable—Lubin says it isn’t so simple.

“Centralization/decentralization is a cyclical dance,” he says. “The Internet and the web decentralized things for much good, and inevitably, smart people figured out ways to capture value and control. It’s time for the next, better foundational iteration.”

Lubin explains that the Internet itself was born decentralized— but as it grew in size, so too did the need for greater efficiency and consolidation. “As more value flows through those pipes, there’s more reason for consolidation,” he says. This occurs not just for technical efficiency but for economic gain as well. “We’re kind of seeing that with Facebook right now,” says Lubin. “Facebook started as a cool system that other people got a bunch of value, either in Harvard forums or in other schools... And eventually, more broadly in society, it has turned into essentially a weapon of mass control—if made available to the wrong people.”

That’s typical of monopolies, he explains. Eventually, once a small group of smart people “figure out loopholes and strategies and work to consolidate power and influence—work to silo information for their own benefit, you end up with what society deems unacceptable.”

“You get a backlash,” he says. “And some grad student or high school student comes along and posits a different sort of architecture with the benefit of the knowledge of what’s gone wrong in the previous architecture—and people get excited about it and start building it.” And the cycle begins anew.

But while efficiencies can often be achieved through consolidation and centralization, it isn’t always the case. “One could imagine decentralized [data] storage being much more efficient than centralized storage in the form of Dropbox, or systems like that,” says Lubin. “If somebody puts together a decentralized protocol that frees [unused] space, or makes it easy to use all that space—and rewards or incentivizes people to make it available—I could imagine a decentralized storage platform that’s much more efficient, and much less expensive, than current systems.”

The same is true of legally enforceable, hybrid-blockchain-based OpenLaw agreements, he says. Building a library of legally enforceable agreements that are properly structured can produce a more efficient, fairer, and much less expensive system. Narayanan’s criticism of smart contracts as “neither smart nor contracts” is “embarrassingly old news,” says Lubin.

“It’s naive to believe that smart contracts are going to be law,” he says, “but we can write laws, which are a social construct, in different technologies.”

Lubin says there’s nothing wrong with memorializing laws, contracts, or even treaties between nations on decentralized systems, when appropriate. In fact, doing so—with the proper infrastructure in place around these agreements for arbitration—is simply “a more efficient technology to build the stuff that we already build. And it’ll enable us to add more richness to agreements.”

As Naranyanan notes, however, appealing to users on philosophical or even cost-savings grounds won’t be enough to win a large number of them over.

No argument from Lubin, who says the fact that it’s going to take attracting developers—ultimately drawing some of the best and brightest—for decentralized platforms to eventually succeed is pretty obvious to him, too. Indeed, developers are already building the next iteration of the web, and decentralization is leading this dance.