Ethereum developers are currently discussing a contentious hard fork core protocol change called Progressive PoW (shortened to ProgPoW). This protocol change will act to reduce the efficiency of ASICs which have started being built for the Ethereum chain in order to protect the dominant GPU miners.

 

Historically, Ethereum has aimed to be a GPU dominated mining chain until the intended switch over to PoS. The design philosophy here is to avoid the perception of miner centralization that ultimately occurs for pure PoW chains with ASICs like Bitcoin, Decred and Litecoin. In theory, if GPUs remain profitable, it encourages miners to be more distributed as consumer grade graphics cards can be used rather than specialist hardware.

 

In order to prevent ASICs from coming to Ethereum, the mining algorithm Ethash uses a high memory (RAM) requirement that reduces the efficiency of ASIC hardware. However as of January 2018, Bitmain released the Antminer E3 which started the manufacture competition by others for Ethereum ASIC miners. What ProgPoW aims to do is essentially brick the ASICs that have been developed and again reduce efficiency relative to GPUs.

 

This has lead to a contentious debate where core Ethereum developers are in favor of ProgPoW whilst many in the community are either indifferent (have no strong opinion), or do not agree with the change.

 

The debate boils down to two key questions:

 

Are ASICs actually a problem for the Ethereum 1.x chain?

 

Is it right for developers to have a say in who is a legitimate miner and who is not (dev superiority)?

IS Etheruem mining centralised?

The first thing we should look at is whether Ethereum mining, even with GPU dominated hardware, is centralised. Consensys performed an analysis for this and came out with two answers:

 

Yes, miner hash power is centralised to the tune of the top five miner pools carrying 84% of the hashrate power. The pie chart below shows the dominant pools compared to individual miners on the network.

No, when we look at the individual miners who feed into those pools. Analyzing the number of unique payout addresses from each pool, it can be seen that there are approximately 285,000 unique miner addresses.

In theory, if pools collude to attack the chain, individual miners are capable of switching to a different honest pool and protecting the chain. That said, there are only 2 or 3 pools that need to collude in order to mount an attack so it remains a possible outcome.

 

In either case, whether ASICs or GPUs are the dominant miner hardware, proof of work always tends towards industrialization. This is simply based on the cost efficiencies of energy (long term contracts), economies of scale and need to upgrade to the latest GPU hardware to retain hash-rate share over time. The average miner will be unable to profitably mine in most parts of the world with a handful of GPUs so it becomes more of a hobby than a genuine game changer for security.

ASIC debate

ASICs are essential component to the incentives and game theory of security systems for Bitcoin and other pure PoW chains that do not aim to be ASIC resistant. The investment by manufacturers and miners in the application specific hardware means that there is CAPEX at risk if the price of the coin were to deteriorate.

 

ASICs can switch to mine any other chain on the same hashing algorithm such as SHA-256 for Bitcoin, BCash, BSV and an other direct forks of the chain. However data has shown that Bitcoin hashrate continues to blitz its clones with less than a few percentage points of hashrate provide to all alternative chains.

GPU miners are more versatile. No doubt there is time, cost and technical requirements associated with switching GPUs to a different algorithm. However the hardware cost is not tethered to a specific algorithm and can also be resold as second hand to recuperate costs. In this sense, GPU miners are closer to mercenaries than soldiers where as ASICs are dedicated to the chain that employs them.

 

Having this as a baseline, are ASICs actually bad for Ethereum?

Checkmates uninformed opinion

In my mind, I think ASICs coming to the chain is a sign of long term investor interest and a guarantee of baseload hashrate. Furthermore, if we assume that Eth 2.0, which use PoS and cuts off all miner income comes to fruition, the mercenary GPU miners will begin their technical switch over before the cut off date, not after. That means hashrate will at the very least have a portion of GPUs and a portion of ASICs.

 

Since the ‘thirdening’ in late Feb 2019 where the Ethereum block reward reduced from 3ETH to 2ETH, the hashrate has been increasing at a very slow and soft rate. Where Bitcoin and Decred are hitting regular hashrate All time Highs this year, Ethereum’s hashrate remains at 62% of ATH which is suggesting that many GPU miners have left and not returned.

 

In this context, Ethereums security is underwhelming, it needs more long term investors to secure the chain. Considering the entire Ethereum ecosystem, DeFi, smart contracts, Dapps and assets are all going to be secured by the 1.x chain until the 2.0 beacon chain starts securing it, it makes a lot of sense to me to let ASICs come in to provide the final line of soldiers to defend its security. Miners are long term investors, every single time the economics of their job changes, it makes them leave, not adapt.

 

That said, I don’t actually understand why ASICs are being a) built or b) purchased for Ethereum assuming PoS is on its way. Seems like sunk capital but the market does what it does.

 

Therefore, I see ProgPoW as a mistake, it is wiser in my mind to let the system and the market sort itself out based on the naturally sound distributed proof of work method for distribution. I am personally a strong PoW advocate and I think the Ethereum developers need more respect for their miners, at the very least until PoS becomes a reality.

Final points

The final point worth raising was our second question, is it ok for developers to have a superiority complex and decision capacity over miners? Should the core devs have access to turn off a miners income stream and brick their investment (ASICs) which sole purpose is to secure their product?

 

I’m sure you can guess my thoughts here…it’s a massive abuse of power and completely centralizing attack vector. Even in the most primitive argument, ProgPoW seems an unnecessary change to the code that has a chance to introduce unknown bugs, doesn’t really solve the centralization problem and shuts down investors who specifically want to secure the chain.

 

From the discussions I see on Twitter (noting it’s a crude form of assessing opinions) I have noted a number of very distinct features of the Ethereum community in relation to this debate:

 

They think very little of miners. They are treated as mercenaries who deserve no special treatment. This has most in the community opposing ProgPoW as it throws a bone to GPU miners. As a trend, I see (and put myself in the middle of) a lot of debates regarding Proof of Work and rarely does an Ethereum supporter have a very sound grasp on the incentives and game theory of PoW. I won’t claim to be an expert, however I have studied it to a higher degree than many and generally receive misplaced opinions from this side of the fence in the debate.

 

They believe mining is wasteful of energy. This is a fair point (but also one of perspective) however the blame is always placed at the feet of miners not…Vitalik…who actually selected PoW as the security mechanism. My go to response is that we as humans waste more energy on meaningless systems like Christmas lights, FX traders with 6 monitors each and military spending than Proof of Work. Side note, everyone loves to talk about how good PoW was at distributing ETH to dilute the ICO…but PoW is also evil…double standards in my eyes.

 

They have no means to effectively coordinate. Twitter and reddit are the general forms of obtaining communication and consensus. I hear a lot of critique of on-chain governance from Ethereum…yet they launch on-chain DAOs to coordinate everything…which is usually accessible only by those with a lot of ETH…who are invariably the same pool of people…who also are the people who are at the moment writing the pushing forward ProgPoW…the developers. On-chain governance is an absolute mistake if doing it in the rear view mirror. It needs to be established from day one with the correct incentives else it is doomed to fail.

 

Developers in Ethereum have way too much say in the chain direction (read: all of it). There is a claim that Ethereum governance is just like Bitcoin’s, an anarchic lack of governance. Yet ProgPoW looks like it is getting pushed forwards despite community opinion or what is (in my opinion) best for the chains security system.

 

Keep an eye on that hashrate chart, especially as 2.0 starts to come live is my take home from this study.

 

If we see evidence of a mass exit of miners before 2.0 is finalizing the 1.x chain, that will be a strong exit signal. Conversely, if we see hash-rate getting a boost and 2.0 development continues to deliver promising results, there may be life in the Ethereum horse yet.