Like Mining and Validation, Proving Needs To Be Decentralized
Web3 adoption is hindered by a lack of privacy.
We don’t always want our online activity to be public, whether when we interact with a social media app or buy a meal, to say nothing of our sensitive financial and medical data.
Imagine a setting on Venmo where everyone can see every purchase — but magnified for all aspects of your digital life.
Dystopian, right?
This goes for businesses too; contracts with suppliers/vendors, capital investments, and even wages would be visible to all, telegraphing intentions and making it much harder to operate and/or compete. For enterprises considering doing business on blockchain-based systems, the transparent reality of an architecture like Ethereum has been a non-starter.
That’s why technologists are keen on zero-knowledge (ZK) cryptography, which allows sensitive information to remain hidden but usable. With ZK cryptography, individuals control what apps and other users see, whether it’s the content of their transactions or the real identity of someone with a blue checkmark on Twitter, for instance — you always have the control to decide what to share and with whom.
Privacy is essential across a range of applications. Blockchains like ZCash have already been able to use zero-knowledge encryption to make financial transactions private and verifiably legitimate. Now we’re on the cusp of entire applications enabled by ZK that are private by default.
So what’s holding us back?
In addition to the lack of privacy, the challenge today facing blockchain-based systems is their ability to scale. We’ve had decentralized applications for years now, but it took decades for the technologies powering the internet to become standardized, let alone efficient enough to scale to the point they are at today.
That said, the pace of innovation in Web3 (and ZK specifically) is accelerating thanks to the infusion of talent and capital into the space.
As blockchains have increasingly moved from proof-of-work to proof-of-stake, we are witnessing a transition from energy-inefficient mining and redundant effort to a more efficient crypto-economic system that relies on verification instead of re-execution. Blockchains that leverage zero-knowledge cryptography for privacy require additional players — provers.
For a quick brush-up, to have private applications on a blockchain using ZK, you need a prover to generate a zero-knowledge proof attesting to the validity of a transaction. For example: someone using a ZK-enabled app could log in without having to send or reveal their password/credentials. Instead, what would happen is that they’d generate a proof that they have the correct credentials to log in to the app, the app would be able to verify the proof is valid, and they’d get signed in.
To add a bit of complexity to this picture, the best way to scale ZK-based systems is generally to have proving be an outsourced job performed by a specialized service provider rather than the responsibility of individual users.
As ZK continues to become adopted, it will be essential to make proving a distinct, decentralized layer of the stack, just like mining and validating.
Unfortunately, most ZK blockchain projects today (like Starkware and most zkEVMs under development) rely on a single, centralized prover. By creating a bottleneck at the proving layer, these projects are unable to enable user privacy and sacrifice the permissionless nature of the blockchain they are building on.
By decentralizing proving and making it a marketplace akin to what we see with validators on proof-of-stake blockchains, we reap multiple benefits. For one, it’s crucial for blockchains to be truly permissionless, which is their killer feature. On top of that, a decentralized proving ecosystem helps accelerate the development of specialized and efficient hardware to address the current computational bottlenecks of ZK systems.
By unleashing market forces with decentralized proving, we help make ZK-enabled applications viable and scalable.
The future of decentralized proving
I’m encouraged that in addition to the decentralized proving layer we’ve already launched at Aleo, others in the ZK space are also leaning into this concept with projects like =nil;’s Proof Market. I expect that this trend will lead to a proving industry with a number of service providers, similar to what’s happened with cloud computing. Already, over 40,000 provers have entered the market, and the competition is just getting started.
This specialization is already happening on both the hardware and the software levels. We’re currently seeing provers using hardware like FPGAs (field-programmable gate arrays) as well as GPUs, some of which were even previously used for mining on Ethereum. Eventually, the growth of this industry will spur chipmakers to build ASICs (application-specific integrated circuits) tailored specifically for efficient ZK proving.
Collaboration between computing hardware giants and the ZK community is also underway and encouraging the development of this technology. Last year’s ZPrize, for instance, brought together companies like AMD and Samsung alongside a number of leading ZK development teams.
The nature of the competition was similar to that of DARPA’s Grand Challenge, and participants raced to optimize the software side of ZK proving on various hardware platforms, from GPUs to FPGAs to mobile phones. The competition yielded average improvements across these hardware categories ranging from 2x to 11x! Better yet, all these optimizations are open-source to help accelerate the adoption of ZK and scaffold the growing ecosystem around ZK.
Zero-knowledge cryptography is going to transform the internet as well as the infrastructure and services built on top. I envision a whole new category of cloud services built around ZK, with proving as an example of one of the most important of those. In the coming years, I can’t wait to see the advent of a private yet personalized web experience enabled by zero-knowledge cryptography.
