You do not often see 300 people waiting for the start of an eth dev meeting where very technical matters are often discussed.
Friday, however, was different because for the first time miners had been invited to take part as well as representatives of eth holders to discuss issuance reduction.
Yet the meeting begins as a gentle affair of minute details about certain Ethereum Improvement Proposals (EIPs) and whether they should be included in the upcoming Metropolis Constantinople hardfork.
But there were signs throughout that this wasn’t quite a work-like eth dev meeting where they catch up with what has been done so far and what needs to be done in the next few days.
Nothing crazy going on in research, says one eth dev seemingly keen to get off the microphone. A baby says wa to the delight of a now sizable crowd. Eth’s ratio against bitcoin starts to fall.
Brian Venturo, who has 3,000 GPU rigs, takes the mic to argue in favor of EIP 1295. He wants to remove completely the coded increase in mining difficulty implemented to incentivize the network’s move to Proof of Stake (PoS), better known as the difficulty bomb.
He further wants base block issuance to remain at 3 eth, but uncle (orphan) rewards to be reduced. That would give a 10% issuance reduction, or an inflation rate of 6% a year, down slightly from the current 7%.
He argues miners are currently incentivized to increase uncle rates because more uncles means more reward. This EIP, he says, would address that misalignment.
The well spoken Venturo often attracted boos by the crowd. Speaking confidently, seemingly very well prepared, and in a very business like manner, you could say he dominated the discussion in this gathering of circa 35 participants.
Carl Larson, in contrast, appeared very nervous. His voice almost shaking, courage he was taking to speak in favor of EIP 858.
That wants to delay the difficulty bomb for another year while reducing issuance to 1 eth per block. Presumably that would have a proportional reduction in uncle rates as otherwise miners would receive twice more eth per orphan than a real block, but the EIP doesn’t quite say so with the delay in the difficulty bomb added to it only recently.
This scruffiness might be because Larson is the founder of ethtrader, rather than the operator of a huge mining farm or an eth protocol dev. For that he is given some leeway, with Larson’s argument focused mainly on the environmental benefits from the issuance reduction.
Venturo counteracted by saying there wouldn’t be much benefit as miners can keep using those GPUs in other coins. Venturo concedes, however, that there might not be as much growth in GPU mining at 1eth per block.
Larson says 65% of mining GPUs are mining on eth. Eth dominates by far, he says. Venturo argues miners use renewable energy anyway. There he gets shut down by his own colleague, Xin Xu, CEO of SparkPool.
Xu says “all Chinese miners are using non-renewable electricity.” So in effect closing that debate, even though Xu did then argue miners are using energy that otherwise wouldn’t have been used by restoring old power stations.
Xu further had a very interesting point to make regarding security. He says it is cheap to 51% attack ethereum, but no one does because there is no market for it. Otherwise said, they’d get far more by pocketing the current $2.5 billion a year in block rewards.
Xu, however, is of course not in favor of reducing it to one eth. Wants instead to keep it at 3 eth per block for the pool’s 300,000 mining rigs.
Just how much 1234 was missing from the debate is shown by our notes not mentioning it once. Afri Schoedon from Parity said he would champion it, but ended up doing no championing in this meeting. It was left to Eric Conner, a long time eth holder, who pointed out that when eth’s issuance was halved last year, the hashrate tripled.
That 1234 EIP is seen as the compromise with it proposing to reduce issuance to 2 eth while delaying the difficulty bomb for a year and a half.
Why a year and a half, why 12 months, why delay it at all, was not discussed even though one can argue that this continued delay leads to less pressure to deliver PoS, just as one can argue eth would continue working fine with block times at 45 seconds or even a minute until Casper PoS.
What was discussed, however, was whether miners would really leave if issuance is reduced to 1 eth or 2 eth. Tim Coulter, founder of dev-tool Truffle, but present in the meeting to represent small to medium miners with 100 GPUs, says he would keep mining if it is reduced to 2 eth and even to 1 eth.
Ethermine had sent a statement saying their miners were ambivalent regarding issuance, with some in favor of 2eth, some ◊1, some wanting to keep it where it is. Thus they took no position on issuance.
Venturo says it wouldn’t be economic for him to mine at 1eth, but can manage at ◊2. He further says he is in favor of a reduction if it is coupled with a Proof of Work (PoW) change.
Afri Schoedon recently highlighted a new algorithm called ProgPow which says: “The design goal of ProgPoW is to have the algorithm’s requirements match what is available on commodity GPUs: If the algorithm were to be implemented on a custom ASIC there should be little opportunity for efficiency gains compared to a commodity GPU.”
Hudson Jameson, an eth dev chairing the meeting, says this wouldn’t be available for Constantinople, but they have now suggested there might be a third fork of Metropolis, in about eight months, which is probably to be called Istanbul. ProgPoW might be implemented in that fork.
One meeting participant says eth asics are on par with top of the line GPUs. The constrain is in memory bandwidth, says another, so gpus are asics, he argues.
Those points might be irrelevant as in eight months Casper does hopefully make its way out, but until then all seem to be in agreement that the difficulty bomb should be delayed, with a reduction in issuance proposed to keep inflation at the level it would have been had the difficulty bomb not been delayed.
That aim of keeping inflation at the same level has not quite been met, with ◊5.5 million, worth some $1.5 billion, mined in addition to what would have been mined had the difficulty bomb been kept.
A reduction to 2 eth or even 1 eth would probably still not keep it in line with the difficulty bomb inflation rate, but it would do so more than at 3eth.
Once Casper, it then goes down to 0.82 eth, and eventually once full PoS, it will be at around 0.22 eth per block, with a long term issuance plan therefore seemingly laid out.
As far as in the short term, the meeting concluded without a decision while the crowd of now 600 was chanting 1234. They are now to meet this Friday where the matter is to be discussed once again.
Schoedon says 2eth is a good compromise. Larson says the cards were stacked up against him regarding a reduction to 1 eth, even though most token holders prefer it. Vitalik Buterin has stayed out of all this, as have a number of other devs, including Vlad Zamfir or Gavin Wood.
The question now is how they will decide, with no one seemingly willing to say it will be 1 or it will be 2. For an answer to that, we’ll have to wait, but it seems unlikely any dev will argue to keep it at three in the next meeting, so there might even be consensus.