In the latest season of HBO’s popular series Silicon Valley, the iconic “CEO” of Pied Piper, Richard Hendrick proposes a stunningly ambitious idea for his startup: rebuild the Internet as a decentralized network that utilizes the computing power of billions of phones in our pockets and spare computers in our living room.
This is in stark contrast to the world we live in today, where more than a third of all Internet traffic goes through the few dozen massive data centers of Amazon Web Services’ across the globe. Recently, that architecture led to a massive Internet meltdown after a regional outage at Amazon’s Virginia data center.
Hendrick’s big idea to decentralize the Internet was obviously inspired by the Ethereum project, a world computer based on complex cryptographic protocols led by now 23 year old russian whiz kid Vitalik Buterin. Recently, we witnessed Ethereum token’s meteoric rise to over $25B in market cap, all happening in less than three years since its launch. If Ethereum was a startup, it would be considered as one of the fastest growing unicorns in history.
Unlike the traditional cloud that can crash or be hacked, Ethereum is often referred to as a “perfect virtual computer”. It is unstoppable, uncensorable, tamper-proof, and impossible to catch malware or a virus. This is achieved by a complex piece of cryptographic protocol that runs over a large network of individual computers across the world; rather than being concentrated in a few data centers. Anyone can join that network and become a “miner” by lending their computation services and earn digital tokens, known as Ether.
Yet, despite its phenomenal growth and serious enterprise backing from world’s leading enterprises and financial institutions from Microsoft, Accenture, and JP Morgan, researchers in the space so far caution that we’re still at the earliest stages of development. Vlad Zamfir, a well-known crypto researcher and core member of Ethereum Foundation put it as follows:
Let’s look at what he meant by not scalable. Take Twitter, a popular and centralized social media service as an example. The site processes hundreds of millions of tweets per day on average. And that’s just a single service. The capacity limit of the entire Ethereum network is currently on a scale of less than a million per day. Yet, adding new computers, aka “nodes”, to the Ethereum network doesn’t help it to scale, due to limitations of its current algorithm design.
Solving for coordination of millions of computers across the Internet while able to achieve scalability is certainly not a trivial task. Many of the brightest minds in mathematics, cryptography, and economics are in a major race to solve this hard problem, including Ethereum’s own next-generation research initiative, codenamed Casper.
Among many contestants, the DFINITY project is an intriguing one with its strong technical underpinnings and a particularly ambitious vision to focus on delivering a “Decentralized Cloud,” rather than automating trust which most blockchain projects tend to focus on.
Supported by DFINITY Stiftung, a not-for-profit headquartered in Zug, Switzerland, the project boasts over $20M in funding (with another main round fundraiser upcoming), and a strong team comprised of top engineers, scientists and economists, many with background from leading organizations such as Stanford, Google, Yale and University of Chicago.
“Our objective is to deliver a public decentralized cloud computing resource, with vastly improved performance – already seeing more than 50x that of Ethereum in its first release this year – with the goal of ultimately achieving infinite capacity.” says its founder Dominic Williams, a British-born serial entrepreneur. Williams’ last multi-million user startup in the massive multiplayer space brought him to the Valley, after fundraising from leading US venture investors.
The ingenious way that DFINITY achieves such performance boost comes partly from a piece of cryptography invented at Stanford, named BLS, standing for its three co-authors: Dan Boneh, Ben Lynn and Hovav Shacham.
Today, Boneh is best known as the head of applied cryptography at Stanford and one of the world’s leading cryptographers. Lynn, his former Ph.D. student, spent the last ten years at Google and recently joined String Labs, the incubator and a lead contributor to DFINITY Project, to work on its core protocol.
“Seeing BLS cryptography been applied to power the next generation decentralized cloud, and potentially used by tens of millions of people makes me incredibly excited as a cryptographer and engineer.” says Lynn, when asked about why he decided to join as a full-time contributor.
A vastly performant decentralized network could replace today’s unnecessarily complicated IT systems on centralized infrastructure. In Williams’ words, “Enterprise IT systems running on this computer will be unstoppable and won't need to involve complex components such as databases, backup and restore systems or Amazon Web Services, allowing costs to be cut by 90% or more by reducing the supporting human capital required.”
The disruption potential doesn’t stop with that, he stated, “A highly performant decentralized cloud will also enable the creation of open source decentralized businesses using self-updating autonomous software systems that may eventually be able to disintermediate and beat out monopolistic organizations such as Uber, eBay, Gmail and others.”
Still, there’s another major hurdle it must overcome. The blockchain computer, while featuring immutability and trust, carries with it a new dimension of challenges that hasn’t been seen in traditional cloud computing. What if the software on blockchain is buggy? What if the funds on it are stolen by malicious hackers? Will these problems be “immutable” as well?
This class of problems are commonly known as governance issues in the blockchain community. Consider the infamous “The DAO” heist in the summer of 2016. A decentralized “venture fund” robot, named the DAO, on Ethereum collected $150M worth of Ether tokens from 20,000 individual anonymous investors in less than 50 days. An amazing feat by all accounts.
Yet only a week later, an anonymous hacker was able to exploit vulnerability in its code that caused a loss of $90M. This ultimately led to a decision to fork Ethereum into a new network that reversed the hack and rescued the user funds.
That decision came from months of endless debates and controversies that haunt it even today. In absence of a formal governance process, people questioned the legitimacy of such a decision in a decentralized system, where software “code” is often believed to be the ultimate law and immutable.
The DFINITY team took apparently a different philosophy:
“While Code is Law is indeed valuable in some cases, we think a different paradigm is needed for mainstream business uses. A world where a twenty-something hacker could happily walk away with multi-million dollars of theft by watching and exploiting software vulnerabilities on the blockchain, is not particularly an appealing one for enterprises or commercial applications. We need a blockchain ‘algorithmic court’ to settle these cases.”
They propose a solution to such problem called the “Blockchain Nervous System”, which adopts an “AI is Law” design. By utilizing a hybrid human and AI governance algorithm, this algorithm could essentially overrule any previous code execution result when deemed necessary.
It also borrows a page from the political governance experience, something known as Liquid Democracy, that allows anyone in the community to delegate decisions to their trusted person, in order to reach high quality and rapid decisions en masse. For example, a DAO-like hack could potentially be reversed in a matter of hours given its non-controversy among community members with a strong following relationship, yet without risk of fragmenting the community with a hard fork.
Yet, with these powerful designs comes inevitably many uncertainties. What kind of proposal would be made? What if controversial proposals are mysteriously passed by the AI? Will “crowd wisdom” bend toward long-term optimizations or short-term market gains?
For now, the DFINITY team seems to confidently stand by this new experimental philosophy, and believes that it introduces a level of governance protocol that is desirable by mainstream enterprises. Leading organizations including Boston Consulting Group seem to concur with this in partnering with the project.
With its recent briefing, the SEC deemed the sale of digital assets the same as selling securities, which means tokens may be subject to the same laws and regulations. While some say this may have a chilling effect on the industry in the short-term, more regulation will benefit the industry in the long term. The market also seems to have already priced in this decision by the SEC, ruling out the worst case scenario of a complete ban of tokens.
“I see this as a positive long-term development. The SEC alluded to the same thing in 2013 and 2014. The real issue is secondary sales. Free movement of tokens is an essential part of the value of the token. So it has to qualify as not being a security which means structuring it as such needs to be front and center from day one,” said Artia Moghbel, DFINITY’s Head of Operations and Communications.
Controversies and obstacles aside, what DFINITY has successfully accomplished has expanded the scope of what blockchain systems are capable of, and offers a glimpse of what the future of the Internet could become. It’s indeed be an interesting time to watch how decentralized computation, crowd wisdom combined with artificial intelligence, could perhaps finally make blockchain the challenger to the Amazon Web Services and the like.