Coming to Consensus: Proof-of-Work vs Proof-of-Stake

When I first got interested in cryptocurrency, I was immediately enamored with the technology and had to dig deeper and learn all I could. I learned about how transactions are written to a global, decentralized ledger, and those transactions are immutable — meaning they can’t be stopped or reversed — which makes them censorship-resistant and very secure.

As a technophile, born into the dial-up days of the internet, I’ve always been one to become obsessed with the way things work. The how, the why, and the way systems can evolve excites me. Little did I know, the world of blockchain, cryptocurrency, and Web3 would provide more information than I knew what do with.

One of the first things I learned is that Bitcoin maintains its decentralization through a process called Proof-of-Work mining. You’ve probably heard that this is not great for the environment, that it’s slow and wasteful; or perhaps you’ve heard that it’s decentralized, egalitarian, and secure. In a way, both are true, and in a way, neither are.

In this article, I want to take a look at what Proof-of-Work is and compare it to another method of securing and decentralizing blockchains, which you may have heard of, called Proof-of-Stake.

So let’s get into it

Proof-of-What?

I’ll start by laying down a little foundation for this discussion, since, for some readers, this may be your introduction to what is admittedly a rather complex subject. As I mentioned in a recent post, Proof-of-Work (PoW) and Proof-of-Stake (PoS) are commonly referred to as consensus mechanisms, however, this is actually a misnomer. These are Sybil resistance mechanisms and are only one-half of the full consensus mechanism employed by a given blockchain.

According to Ethereum.org, Sybil resistance "measures how a protocol fares against a Sybil attack. Sybil attacks are when one user or group pretends to be many users. Resistance to this type of attack is essential for a decentralized blockchain and enables miners and validators to be rewarded equally based on resources put in. Proof-of-work and proof-of-stake protect against this by making users expend a lot of energy or put up a lot of collateral. These protections are an economic deterrent to Sybil attacks."

With Sybil resistance being at the core of securing a blockchain, it’s no wonder that tensions run high over which mechanism is chosen. Interestingly though, many people don’t fully understand how these work, how they differ, and how similar they are to one another.

Below I’ll briefly explain how Proof-of-Work… works, how Proof-of-Stake differs, and finally, I’ll make the case that one has more upside. But which?

PoW: How Does It Work?

At the core of the Bitcoin, Ethereum, Flux (and many other) networks is a mechanism called Nakamoto Consensus. First described in a 2008 whitepaper by a pseudonymous cypherpunk (or group of cypherpunks) called Satoshi Nakamoto, this combination of Proof-of-Work Sybil resistance and the longest chain rule for maintaining a canonical blockchain history was — and is — groundbreaking.

Jargon aside, what this looks like, in the simplest of terms, is the following:

Using specialized hardware like computer graphics cards (GPUs) or application-specific integrated circuits (ASICs), networks like Bitcoin require those who validate the transactions to provide proof that they have done a difficult and energy-intensive task.

This function is called mining, and miners are engaged in a race to find an extremely random number called a nonce.

The nonce is the missing piece of a cryptographic puzzle that is holding up the current block from being finalized so that the next block can become available. Given the pace at which transactions are intended to occur, the first miner to solve the puzzle receives a block reward: a predetermined payout from programmatic inflation doled out in the respective block.

Those with more computing power (referred to as hashrate) will be more likely to win block rewards. In order to attain that computing power, miners–and more often mining companies or pools–will need to purchase more hardware and expend more energy.

This is where the energy debate usually begins.

Environmentally minded folks worry that Proof-of-Work’s energy usage begets a huge carbon footprint. After all, much of the world is still getting its power from fossil fuels.

Interestingly, the data does not seem to support this argument, as many mining operations set up shop where energy is cheapest: near hydro-electric dams, solar and wind farms, and other renewables. This also demonstrates that Proof-of-Work is agnostic to the kind of energy used. And, interestingly, the idea that Proof-of-Work can serve as a means of monetizing energy itself creates some unique opportunities. For example, solar, wind, geothermal, and tidal operations typically produce more energy than they can use or store, creating what is known as stranded energy.

This energy could be stored (as battery technology improves), but even then, there may be more value in diverting some or all of this excess to mining. The reason? Building renewable energy infrastructure usually requires subsidies from governments, which slows the rate at which it’s rolled out. Whereas operations that can monetize stranded energy could hypothetically fund themselves, perhaps eliminating the need for subsidy altogether.

This is just one of many interesting ideas, all of which are worth studying further. Here’s an interesting paper from Square & Ark Invest that’s worth checking out on this subject. Now to Proof-of-Stake

What’s At Stake with PoS?

As we’ve seen above, Proof-of-Work is, at its core, about trading capital (investments in equipment and property; ongoing energy costs) for the opportunity to validate blocks and thus earn rewards. This is done in a roundabout way that creates numerous externalities, some positive, as discussed above, and some negative, which we’ll come back to later.

Proof-of-Stake is built around that assessment. Its proponents assert that if all that’s truly happening in a PoW system is that someone is putting their capital at stake by investing in equipment, property, and energy, then why not just stake the capital itself?

Since it was first implemented by Peercoin in 2012, Proof-of-Stake (PoS) has grown to become the most prominently used method of Sybil resistance for cryptocurrencies and other blockchain-based projects. Outside of Bitcoin, Ethereum, and Litecoin, nearly all of the most well-known blockchains (Polygon, Avalanche, Tezos, Cardano, Solana) use Proof-of-Stake.

Furthermore, Ethereum itself is fast approaching a long-awaited conversion from PoW to PoS.

In essence, running an Ethereum Proof-of-Stake validator requires either some basic computer hardware (apparently a Raspberry Pi is sufficient) and either 32 $ETH (around $80,000 at today’s prices), staked in a smart contract; or any amount of $ETH deposited into a staking pool—I’ll come back to that in a future post.

Proof-of-Stake validators are incentivized to participate honestly to earn rewards. They are disincentivized from attacking the system, in most cases, through a penalty called Slashing. This is an automated process built into the staking protocol that burns a portion of the staked capital of those who violate any of a small set of clearly defined rules.

In the case of Ethereum’s Casper FFG system, malicious actors will only have three chances before they’ve had their entire stake slashed. This method of fraud/censorship prevention reduces the need to fork the chain in the event of an attack. However, that remains a completely viable option in the event of a sustained 51% attack.

Proof-of-Work or Proof-of-Stake - who wins?

While the simplest appeal, to the arguably needless waste of energy by Proof-of-Work, is often the one made as justification for why Proof-of-Stake is better, that’s not the narrative I intend to follow.

Proof-of-Work can work and could be ideal for some blockchains. Needless to say, it provides a way to monetize stranded energy, subsidize green energy, and at the same time decentralize blockchain computation. However, there are several areas where I believe PoS outperforms PoW, which we’ll go over below.

Note: While I’ve formulated my opinion through countless hours of absorbing information from whitepapers, interviews, and reflection on the subject, much of what I’ve come to believe overlaps with the points made by Charles Hoskinson in from an interview with Lex Fridman. As such, I recommend watching it as a supplement to this article.

Egalitarian Participation

While some may argue that the barrier to becoming an Ethereum validator ($80,000 at today’s cost) is higher than becoming a Bitcoin miner (entry-level ASICs can be purchased for around $5000), this is a bit misleading. As I mentioned above, those winning the majority of the block rewards are those with the most hashpower, and since increasing hashpower benefits from economies of scale, larger operations disproportionately benefit. This creates a feedback loop that adds to centralization.

Similarly, those with access to cheap energy, like governments, for instance, can easily create major mining operations. On a small scale, this seems like a good thing. El Salvador, for example, is using its Bitcoin profits (mined renewably from a geothermal operation atop a volcano) to provide public goods to its citizens. But imagine, for example, several countries with large amounts of energy at their disposal, like Russia, China, Saudi Arabia joining forces to exercise 51% attacks on the system and censor transactions from their adversaries.

This brings me to my next point…

Security

Not all Proof-of-Work is run by ASICs in densely packed warehouses. It is entirely possible for numerous small participants to contribute to the security of the network. In a hypothetical scenario where every citizen of the world had one or more miners, the level of decentralization would create tremendous security. However, in the current system, and in some ways even in that hypothetical system, large mining operations would still be easy targets.

For example, if tomorrow, a non-economic actor (someone who cares more about destroying a blockchain network than the economic costs they will incur; like a government), decided to attempt to attack the Bitcoin network, finding large scale operations and thus the majority of hashpower is very easy due to the footprint left by their energy consumption.

In a Proof-of-Stake system, validators are effectively invisible. They have blockchain addresses that can be traced, but their energy footprint is no more than that of an ordinary computer. It can quickly and easily be relocated across town, across the country, and across borders.

This brings me to my next point.

Hardware

While everyone focuses on the energy use of running blockchains, they miss a far more important point: the hardware used has a shelf life. Over time it gets worn down and becomes less profitable. In some cases, it's refurbished and resold to amateur miners, which is a wonderful way to extend its life. But nonetheless, the final resting place for all of this equipment is the landfill. Bitcoin at present serves only a tiny portion of population, but requires tremendous amounts of specialized hardware, running at high temperatures 24 hours per day.

As it scales, the potential for this eWaste to get out of hand is inarguable.

Mining hardware also has a centralization problem. Popular ASICs are made by only a small handful of companies that could themselves act maliciously, or more likely become targets for state interference. Graphics cards companies, like NVIDIA have already shown a predisposition for bricking their own cards. And of course, they do this for a reason. Demand for graphics cards to build mining rigs has altered the markets and made buying a card for gaming or professional applications overwhelmingly costly.

This brings me to my final thought on the externalities of using hash power for mining.

Why?

If you agree with the premise that both Proof-of-Work and Proof-of-Stake are equally dependant on capital, and thus the hardware expended arguably does not add sufficient value to the process, then wouldn’t it be more ethical to pool that hashpower to do other valuable things like open-source medical research, or modeling climate change solutions, or at the very least using all those graphics cards to democratize access to 3D renderings, like what Render Network is aiming to do?

In my humble opinion, it would. But that’s just me.

Hopefully, I’ve provided a little food for thought. I see potential use cases for both mechanisms. What about you?

Also Published here