Reducing block subsidies incentivizes miner-induced chain reorganizations, since the risk/reward of looking backward in the chain to republish recent expensive transactions to steal their high network fees starts to become more favorable than looking strictly forward to the mempool for legitimate operation.
Furthermore, mining related EIPs have been systematically deprioritized for a few years so that developers can focus on strategic PoS tasks, so this EIP should also be similarly deprioritized. If adopted, this EIP should be pushed after the existing mining related needs, like fixing known Ethash vulnerabilities (details of which are documented in the various ProgPoW analyses).
Ridiculous proposal at the end of PoW.
Dishonest to say mining will not be affected or security either.
Why should we compare inflation rates to BTC?
What is the problems for miners to profit if the whole chain integrity rellies on them?
We won’t be worrying about inflation if DeFi craze combined with lower processing power in the chain start making bottlenecks.
It’s time to have the best version of ETH 1.0, and not jeopardize it just before the giant leap ETH 2.0 will mean.
This EIP is at least irresponsible.
If EIP-1559 is achieved before this proposal, AND IF the fee-burning is of a non-trivial amount that will reduce inflation to below 0.5ETH, then there will be no urgent need to activate this EIP.
We do not think this is likely to occur in the short or medium term. 1559 will not be activated on the next hard fork, so it is as least 6-12 months away. Then after it is activated, 1559 requires end user wallets to update their fee handling. Then, after both of those things, the fee burning needs to be large enough to reduce the inflation to a sufficient amount.
We do not believe any of these things are likely to occur in the short or medium term, and therefore we would move forward with this EIP to reduce inflation and to reduce over-paying for security.
If EIP-1559 is achieved after this proposal, nothing would change from a technical perspective, there would simply be less deflation of the money supply by 1.5ETH per block on average. We prefer this to scenario #1 (which we feel is unlikely to occur any time soon), but would happily accept both.
We would welcome any proposal that reduces the block reward, even if it includes ProgPoW or other changes. However, that is not the focus of this EIP. If you would like to fork this EIP and add support for ProgPow, we would happily support your proposal if it can be shown there are no technical vulnerabilities in the new mining algorithm. We will leave it to someone else to propose that EIP.
Your fiat earnings are not being cut. Miners were mining 2ETH@$75 = $150 in December 2018. Miners will be earning 0.5ETH@$350 = $175 if this change happened now. And they will be earning 0.5ETH@$490 = $245 if it returns to its August highs.
We are simply using Bitcoin, among other blockchains as evidence and reference that Ethereum is paying too much to block validators. We don’t make any other claims regarding the age or the blockchains or other features that one may have over another.
Miners, in our view, are security-for-hire service providers. There is no social contract, nor has there ever been, that miners should expect Ethereum block rewards to remain fixed indefinitely. This has been the case since 2016 when the rewards were cut for the first time. We fundamentally believe that economics of the larger ecosystem are more important than that of the miners. Without a high price for ETH, miners would not be earning any money in the first place. To suggest that miners economics must be first and foremost is like saying that “the tail should wag the dog”.
Good point. And just a few days ago, blocks has up to 11ETH in fees, with 2ETH in block rewards. Even if the block reward was set to ZERO, miners would be very happy! With all the continuing DeFi tx that we see occurring, we don’t expect fees to drop any time soon.
We agree, and do not want to dismantle anything. We want to make it more efficient.
What if the majority of miners won’t support the switch to ETH2.0 PoS?
Economics will always dictate what miners do. The non-canon fork will not be called Ethereum, it will be called ETH Classic 2, or some other name. Exchanges will continue to call the Core-dev-sanctioned chain “Ethereum”. It will have more value, since convincing everyone to switch to your new fork is hard. This was tried with Ethereum Classic, and we can see the results there all to clearly.
There is no precedent for miners to fork the chain on this issue. They did not do this on the previous two hard forks that reduced the block reward. Also, ETH holders control $45B in wealth. This includes exchanges and all the other market participants. They will not side with miners. Instead, they would side with a chain that will have less inflation and therefore less sell pressure, and therefore increase their wealth.
In short, we are not concerned with this scenario. Miners will happily mine on the 0.5ETH chain because the fiat rewards will be much higher than the forked chain. So , by all means, fork the chain and mine on Ethereum Classic 2 if you so desire. We encourage competition. ETH holders would have tokens on both chains, so they would not take any net loss either.
With all due respect, “coming quite soon” seems to mean “slightly better than the pace of a government IT department”
Looking at Ethereum’s original road map, its been almost 3 years since the first version of PoS was supposed to have been already launched. If we had one wei for every time we heard “quite soon”, we would be rich enough to retire. We believe, based on the pace that has been taken so far, it will be a number of years before the PoW chain is fully de-activated.
In the meantime, if Ethereum continues to make progress, the price could go back to 2018 highs of $1,500, and miners would be earning exorbitant, inefficient, extortion-rate rewards. This is not fair to ETH holders.
Last week, during the Uniswap token launch, we saw fees at 10-12 ETH per block. That is 6 times higher than the reward of 2ETH. In your calculations, what is the FEE/REWARD ratio that you feel would be ‘unsafe’? Are you suggesting that If fees remain at 10ETH per block, should we instead be increasing the block reward? If so, what should it be increased to?
Keep in mind that miners make their hardware purchasing decisions in FIAT terms, not ETH terms. This is because all their expenses must be paid in FIAT. (hardware, electricity, real estate, labor, etc). Idle hardware will occur any time a miner would be losing FIAT money by running their hardware. This does not only happen in the scenario you describe. This can, and does, happen in reward-dominant chains as well. And it happens because the price of those tokens drop too low, which in turn reduces the FIAT rewards for the miners.
Let’s take your example for a moment. No rational economic actor would suddenly buy 5x more mining hardware because of the short term fee spikes we have seen from DeFi apps. They will only do so after a long-term period of high fees. If the fees are high for a long-term period, that would mean that there is high demand for Ethereum transactions. Presumably, many of those transactions would be through smart contract DApps similar to what we have seen over the past weeks (DeFi, lending, DEXs, etc). If there is high demand for ETH transactions, all else equal, that would result in a higher demand for ETH itself, which would result in a higher price. High fee/high demand periods also attract large amounts of stable-coins and other assets to be issued on Ethereum. Those are sticky economic drivers that have a positive feedback loop effect that does not unwind very easily or quickly. The higher price due to these actions would then make up for the low-fee periods. Miners would still earn enough in FIAT to avoid idle hardware.
In the disaster scenario you might argue for, where there is a protracted and long low-fee period and a large price drop in ETH, we could see idle hardware. However, if this were to occur, we believe Ethereum would have bigger problems on its hands. Such a scenario would likely mean that Ethereum did not innovate fast enough, and other competitors started capturing its market share.
We could see this unfortunately happening if there are continued delays deploying level 1 scaling. In the disaster scenario you might allude to, Ethereum would no longer be in demand, and miners will have nothing to mine because ETH will have little or no value. We rely on you and the core devs to deploy ETH 2.0. In the meantime, if we drop the block reward to efficient levels, there will be less inflation and supply, and that will (all else equal) lead to a higher price. A higher price means that more capital, more investment, more liquidity, more talent, and more building happens on Ethereum.
We fully support EIP-1559, and hope it will activate in the next hard fork.
Not 5x, but that translates to literally millions of GPU’s.
Having the most secure GPU-centric blockchain on the planet seems fair, no? Through the highest of the highs and the lowest of the lows, those GPU’s keep cranking on because investments into the security of the blockchain were made.
I hope you realize that December 2018 was the beginning of what we call “crypto winter” where we mined and sold pretty much everything just to keep our lights on. Many mining ops were in debt, and many went bankrupt. I pay under 3 cents/kwh for power and we barely hung on. Reducing our fiat earnings back to this level will put the entire industry at risk again, but this time will be worse because there is no hope for better earnings in the future. See my next point, as your fiat earnings calc also ignores difficulty as a hindrance for earnings for miners.
This is not a fair comparison due to another issue unaddressed by ETH that does not affect Bitcoin: ASIC’s. There are no Bitcoin ASIC’s that are far more profitable (to our knowledge) that makes the public BTC ASIC’s weak by comparison. This is another uphill battle honest GPU miners must face today. We have no idea how much of the network is ASIC’s, best guesses put it at 50%. While the upgrade with Bitcoin ASIC’s v. Ethereum ASIC’s are not as pronounced, it dramatically affects the bottom line for GPU miners, creates more centralization and security risk in the network. Solutions have been rejected by the (mostly non mining) community.
I agree with your sentiment here, and miners know the community’s feelings towards them all too well. But you are promulgating a change to an entire industry with an incomplete picture and drawing incorrect conclusions regarding the health of the Ethereum mining ecosystem. You are pricing out good actors and leaving only ASIC chips on the network. You cannot reduce inflation and leave ASIC’s on the network. This is not good for Ethereum, is at odds with the yellowpaper, and creates security risks.
Unfortunately, and you should know just as well as anyone, that layer2 solutions has the potential to all but erase any larger earnings miners have enjoyed recently as a result of DeFi. Telling miners they should be “happy” with an inconsistent revenue stream that could be erased through a myriad of other soon-to-be-deployed solutions feels very disingenuous to me.
Then it sounds like you would agree we should have a discussion around lowering emissions, at the very least commensurate with emissions lowered in the past, as part of a discussion around removing ASIC’s off the network AFTER Serenity is deployed.
Not to mention EIP1559 + this proposal could eliminate mining rewards pretty much entirely for miners, and would leave mining to be approachable to a handful of chip makers with free hydro power in China.
You mean like the last 3 years? This is entirely possible, in fact, one would argue that crypto is cyclical, and will happen again.
Then this EIP should be a non starter, as EIP 1559 addresses the concerns this EIP addresses more eloquently. While I am biased, and disagree with EIP 1559 in that it is radical and destroys some incentive for miners, it at the very least ensures that mining rewards are stable, and wouldn’t force miners to rely on the latest DeFi meme to pay the power bills.
This is why we are illustrating that vitalik’s point about high fees should not be a large concern unless they are sustained over a long period. And if that happens, EIP-1559 would solve the inflation issue sufficiently; thus eliminating an urgent need for a block-reward reduction.
However, that is not the case now. We believe it will take more than 2 years for 1559 and PoS to be fully deployed.
Here’s the screenshot on a proper time-period for real context.
We are still far below the peak in 2018. Furthermore, the increase in hashrate is more likely attributed to the price increase from $100 to $490, not a few days of high fees.
Finally, prior research has shown that the Ethereum blockchain is at least 40% ASICs, but likely more by now. We support a switch to ProgPow, since we prefer GPU miners, but the ASIC lobby unfortunately killed that idea.
My guess is that it would be best to leave it as is and let miners get their hardware paid back. Many miners lost a lot of money during the cryptowinter. 4x block reward reduction to 0.5 ETH is a bit overkill; it would be a big surprise for new guys that entered the mining for the first time.
You’re wrong. I watch this daily. The recent bumps in hashrate are entirely attributed to news in DeFi and increasing transactions… which has been driving the price up.
Just so you and I are clear, who is this “we” you keep referring to?
The hash rate dropped by about 50% when we were at the lows, but the price dropped by 93%. Therefore, there were still many miners who had lower costs than you and were able to turn a profit or break even. Your individual situation does not change the fact that the mining hashrate was still robust. Secondly, GPUs are faster and more efficient now, so there is some minor bias as to what the actual amount of work done is (which is more imporant than hash rate).
Keep in mind that miners don’t increase the hashrate by pressing the “increase hashrate” button. GPUs are not shipped instantly, and wiring the things together is not instant too. 250 TH/s is not the peak.
Innosilicon ASICs are not much profitable than regular GPUs and are not universal too. I think that 40% estimate is too high.
