All Core Devs - Execution #45

Transcript

[Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music] [Music]

[Music] [Music] you

okay I believe we are live hello everybody welcome to the meeting we're gonna go ahead and jump right into the agenda if I can find the link all right we're gonna start out with an update from testing Dmitry if you want to start with that and then we'll get anybody else who wants to chime in like Martin or others okay hello everybody and I started to contribute some new test cases on create tool and we are currently discussing how to process this no proof blockchain tests these are

other client developers and this is one thing I want to ask from you is this so last time I introduced roadmap for the test cases and I asked to put some comments is a your ideas and test cases so I would like to ask if it's working or not because I haven't saw any comments on Google document about that and then I also like to tell you that the RAM winning we need test cases to cover they would not assume that we could go to the release anytime soon I'd like to make a

better test coverage for all of them all those scenarios that you have is a new up course and stuff but we don't want to end up is some test case and uncovered and potentially leading to consensus ischium so xcode cache and create two test cases already available and everybody could try it now on test three boy is around pull requests and i'm going to be measuring first part of the test case as soon into half as well so that i could pull it from the master branch me okay so Demetri let me make sure I have

this correct so you just talked about making some cases like create two but also that you're gonna need more time to create test cases so we should maybe adjust the schedule of releasing before Devlin - yes it might be that we won't make it until death contour because I don't want to rush that that's what I mean my point okay yeah it's looking more and more like it's gonna be pushed beyond DEFCON and we'll hopefully be able to get an idea about when block times are going to reach crazy levels so that we can try to get a a hard fork in before that time so thanks for the

update and we'll we'll adjust the schedule accordingly based on testing and other things okay Martin or anyone else do you have a testing updates yeah we're doing some revamp of the post testing stuff are currently in the works of deploying and preparing the production environments to run the Fosters yeah there's some additional bits and pieces to go in but hopeful we can start pretty soon okay great we're gonna run through client updates real quick keep it really brief so we can just run through them we'll start

with parody Fred major updates we've discussed some issues from the last call but I think they're on the agenda for later so I can bring it up okay great guess Congrats on your release and you have any updates thank you so as for the update just really quickly we did a huge rate major minor rewrite this week basically we have released a lot a lot of fancy features just mining on concurrent and multiple blocks trimming uncles and transactions while my his ongoing push notifications local account prioritizations and and similar

things we a huge shout out to Peter Peter Fechter because he was kind of asking and directing us towards the path of preparing this release it was really really helpful they're kind of really proud of this minor release since we released we did notice one instability and one feature that we missed so we'll probably push out another release on Monday just to to have the whole thing completed as a somehow to put a dot on the undermining release but I guess that was mostly the highlight of our work in the previous few weeks okay thank you laughs I don't believe a lot of has

change for these two weeks in terms of Constantinople hard work we mostly did some tweaks for for it wasn't testing it okay wonderful so let me see here harmony etherium jay i nothing special for us bug fixes developer API improvements and working on considering uppercut for implementation thank you thank you Trinity we're one eeep away from having

all the constantinople stuff done and as of like a week ago we confirm that Trinity does now sync the full main net and keeps up with the chain which is pretty cool great next we have anyone from aetherium j/s here all right what about Pegasus Pantheon a mount from Pegasus here so we're able to sink through block 4.85 No that's a full archive stink we're ramping up json-rpc were able to mine but a lot of our time is really been spent on stability at storage and networking layers and we have no updates

about Constantinople support at this time okay great thank you I don't think Alexi is here for turbo gas is there any clients that I missed yes it's eager from men ah yes great hey guys it's my first time here just a quick update Simona is a new client in Alexia language so we have it's a new clients who have were not full-featured bypassing 100% of vehicles and all frontier from steady 8151 t8 and now we have 67% of byzantium test passing and

we expect to have all test passing and three weeks thank you thank you I have a quick update for Nimbus as well from status all right cool so we've been up to our EVM is is nearing nearing stability so we've been exploring using it through EMC for example would guess running the chain but using our AVM now which is pretty exciting we're working on like clients and imperil threats we have a track going for the beacon Jay that's it

okay great II was them as Lane or anyone here that can speak to e wasum Alex do you want to do a quick update yeah I was trying to unlock - Mike yeah we just recently had a too long week to week long sprint and he wasn't and we have at this moment running t-test and running with Aleph and gift we haven't run them together yet but independently they you're able to run it wasn't correctly and we also have a block Explorer for

them and they're able to deploy contracts metered our next focuses to work on language support and a toolkit to to interact with there was M contracts okay thank you uh I felt like there was one more Oh research does anyone here from research maybe Danny who can give an update yeah nothing crazy going on in research a lot of different efforts working on the RNG for the beacon chain a lot of implementations of the beacon chain and

as that I've been doing that the spec has really been tightening up and there's an effort going on for a formal analysis and verification of the apocalypse Casper which is a modification of F of G that you can see on each research otherwise you can follow a lots of little discussions on a three search right now okay thank you I think that's everyone we'll move on to agenda item 4 which is Constantinople we'll go through the EIP is in order EW 1014 I have that as a IP 1014 issues but I'll be honest I forgot what issues that those were so let me look up from

the previous is that's going to create - I believe that's going to create - and I think we solved our problems last time let's see yeah I think it's it's mine item so I just only proposed a change in the naming of the hash function to replace shaft we with cataract - five six full clarity okay wonderful that sounds good to me you can find a link to request somewhere in the comments below the agenda I see the also it looks like someone from l4 and

someone from spank chain also replied to the skinny create - I think Lane reached out to them is that right lane yep I did reach out to both of those organizations and they both sort of caught up the last I heard from the l4 crew was that they're pretty comfortable with the proposed changes I mean from spank chain actually did post some questions I think to the EIP but I think he's actually still looking for an answer - so we should follow up on that okay until then we'll continue on with skinny create - as is with obviously the changes from Pavel being taken into consideration

because it sounds like a good idea okay I think the formerly the process is like Vitalik has to merge it and that may never happen so okay so I'm good on our corridor okay thanks Danny yeah if you can if you can get him to do that that'd be great Jason do you have something oh there's just a baby I love it okay cool he's muted now all right EW 1218 which is simpler blockhouse

refactoring we're just gonna drop that because we couldn't find anyone who really wanted to do it yep can hear you great [Music] [Music]

[Music] [Music] [Music] okay Martin did you have a comment no I was gonna say that Schultz wanted to jump in this Oh okay sounds good did I felt we had

problems with this before can you and comment on that we looked at this a parity and I think it's a good improvement currently we have all these part coded CC checkpoints and this change may not be able to remove them entirely but [Music] I think Fred cata Kenny one here Fred hey Fred we can't hear you at all I

don't know if you can hear us rats on the train you might be just cut off ah can you speak ah free for the changes at all or should we just wait for Fred to come back I'm sorry I was sick and I'm still catching up a beach have to wait for crepes okay no problem it sounds like he it's a positive meaning it seems okay so that's good we are talking about this so that's the skinny version right Frederick is on

chat by the way he says what would help in the combo but main question is if it's included for Constantinople or not who does anyone would not want to include it for Constantinople okay I don't see a issue with including it then especially if we're gonna be shifting the timeline anyway due to testing Demetri I remember you having some issue with adding it I think am I wrong hey Demetri can he hear us from the

testing team yeah yeah yeah of Kimmi yeah I was asking you did you have any problems with 12-18 or anything with blogcast refactoring cuz i remember you saying we shouldn't do it at one time oh yeah because now right test was this would be like another topic creating the test specifically for the new blog hash all the reasons okay but if in the case that you get more help because it sounds like there's an effort going on behind the scenes to get more help but this be something that can be accomplished no giving that this help you also request some time to start working together and

like introduce people to how things are done and put in concert okay like if you introduce new people a couple of months ago then yes but not not at this point in time I guess okay interesting cuz cuz if you introduce new people at this point in time unless starting doing new tests all of a sudden it will be rushing again and we have like three months before the release no longer do we have three months before the release necessarily it can potentially be longer so if it is longer of course yes I'm welcome to have new people working in

the test okay I think we can make time worst case scenario and this is just an idea so people can comment on this there's gonna be discussions about prog pal which is a replacement for proof of work so if it's like really hard to the change is bundled for like all of these e IPS bundled for constantinople we can have a whole second hard work after if we want this is another issue to discuss because if we delay the time we would want to implement more features to this particular constantinople fork and we should discuss but it's good to

have many changes in one fork oh it's better to have less changes in many hard Forks I remember suppose discussion was yes there was and I think things have changed because at the time we had that discussion there weren't that many IPS and now there are more a IPS to handle so I think the discussion is shifting now does anyone have any idea or comments on that especially if we add Mori IPS having two hard Forks not necessarily constantinople one and two but just constantinople with the e IPS we have and then the second one for like 1218 or any of the e IPS that don't get

done in time I would much prefer to stay on schedule rather than to increase scope and push timelines out since we can always do subsequent hard Forks down the road unless there is a specific reason to fast-track getting something into the heart board which I'm not seeing okay do we have other comments well just to support the other side of the spectrum I think it's kind of important to realize that it's Constantinople I think we wanted to do discuss a simple fork at the beginning of the year so the fact that we're at the end of the year and we

didn't do it yet speaks volumes of how well we organize hard Forks basically what I want to say is that it's fine to postpone certain features but we should be aware that if we postpone it then realistically it won't happen for another year I mean we could surprise ourselves but our track record isn't release Peter are you saying it's better to drop features in order to make the date I mean basically what I'm saying is that the it's a fine line so we can drop features but anything we drop will probably not get included for a long

time then you prioritize yeah exactly at some point you have to ship yeah a logical fallacy at play there because by saying things take a long time it's if we don't include it now then it'll be a really long time we're just perpetuating that same thing which is that it takes us we once we've organized our parts where I had much rather focus on getting our process for getting these things out on a tighter timeline so potentially pushing these

things out his motivations for us to to do this on a more realistic and accurate timeline for their subsequent hard work almost all of our teams operate on continuous integration so we're always ready to ship we've got a lot of features that are really quite independent and get rolled into code bases independently but we insist on acting like a big corporation rolling out releases with the ton of features all at once and then being unable to roll it out because we can't get all of the features ready to go at the same

time and I don't know why we have to do it that way except that we're afraid of making hard Forks [Music] hmmm interesting perspective I think that it's very possible for us to tighten up the schedule and to have a hard fork that doesn't take a whole year to produce but that's just my opinion so the problem with the hard Forks it's not only the implementation I mean one of the big problems is that it's difficult to reach these specifications that's that's really big time think I

agree I think normally efficient in other ways what was that comment was it Peter would it would it help setting a date for the next hard fork after this one as a way to help people God decided like the consequence of delaying a particular feature for the next hard fork maybe yeah at last DEFCON we have we sat in a room a lot of people and thoughts discussed and came upon the model that maybe we

should do hard Forks every eight months or so it was kind of generally agreed upon didn't really work out though I personally think that makes a lot of sense we should be basing we should be scheduling hard for something what it was ready yeah I agree with that I agree okay well if we commit to scheduling hard Forks I don't see why we shouldn't drop a few e ip's and then schedule them for the next hard fork and when I say drop the idea is only the ones that are

like really hard to do testing for or something like 12 18 right now that raises the question of if we push I don't remember the even numbers here but the block - and I believe create zoo was going to delay our current Constantinople timeline if we push both of those into another four feet month belt ext Co - what - which one needs can't meet you know oh yeah net gas meter inks the other one sorry I there

was a third one yep exactly so three total AIP 145 bitwise shifting EW 1052 EXT code hash an EIP 1283 net gas metering with s store without door to dirty maps so Demetri those three e ip's I just named would it be possible to get the implementation and testing of those complete in your opinion before Def Con for metering the idea sounds and then it's bitwise shifting and ext - already done I think he mentioned there already

done xt code hash is not yet fully covered so there is a test road map and i actually insist for you to look at it and maybe good more TIF cases into the road map and same for B - i-15 in a sense like a lot of sense the uhé he wrote test coverage for BTS 15 in the beginning of spring so mission Justin should also reduce this road map that he covered and maybe put some more test cases for bid 550 okay we can have more people help with that if anyone on the call just like

that if they can help with some adding some test cases that would help a lot adding the test cases I not necessarily mean writing the test case and executing it just putting this idea what was the test guy should do and then I write the actual test sounds good okay so Demetri just to repeat my question if you have any idea and if you don't this is okay but if we use the three EIP as I mentioned bit Y shifting EXT code hash and net gas metering without dirty maps is that something that we can reliably

get test for in the next month to then do a month of testing before releasing yes what was create to create to is that's something that we can also do before Def Con it's kind of obtained I already started writing test for create - and I saw the commits merged in CPP year so I saw it was already accepted same condition okay then that could be on there as well so that would actually be all of them that we've already decided on so really the only one we'd be delaying is the one that Zolt was talking about which is 1218 welcome the

block hash yes so 1218 would be delayed and then if we decide on prog pal that would have to be delayed - because there's no way we're getting that into Constantinople with the timeline we currently have however we can do instead of doing an eight month timeline since we've already started on some of these we can do a six month timeline maybe anyone like short little accelerated okay even though we have one thing right now there will be other each to show up to that okay that sounds good alright so in that

case what do other people have opinions on an eight month timeline vs. a six month timeline I'm kind of on a high for now I think that's reasonable I think that should be discussed in the context of pushing the difficulty gone back and so for shooting for a twenty nine does that push us on like a 12 month difficulty bomb to kind of force the hands and so if you delay the eight then you have to at least pork by the twelve is that reasonable rather than like a 16 month or yeah eighteen months it will t bomb that sounds reasonable to me sixty percent buffer time to four when

things don't go as expected but still forces us to make an action in a timely manner great Peter did you evah then I saw you and music only wanted to say that the reason why I would also really oppose a six-month timeline is that basically if the time of if the if the four frequencies are too large and there's too little time in between and that kinds that kind of puts a huge pressure on currently at least on the parity and the gas themes because not only do we have to do the whole fork

stuff but we actually do also have to maintain the network and to ensure so it takes a ton of effort to just to maintain the current network and to keep it running and to do the minor tweaks so it's for us it's a bit more work currently and and you don't really want to put even more strain on the themes so your argument is for the eight months is that right yes yes exactly great okay we haven't even well with emergencies wouldn't be bad yeah I agree the date your date for eight will keep

showing up at different times and it'll always be when's the next one when's the next one instead of like you know it's July here's your fork yep exactly okay I'm comfortable with eight months there will be emergencies there will be flat-out emergencies so that is coming soon or then does this mean that it would be eight months after Constantinople oh the current plan

yeah no it's not gonna - it'll be a whole new name yeah the name has to be Istanbul I think okay it might be a stained bull everyone's gonna have to learn how to spell that one too okay let's talk more about this next week reason being we need to get some time in for the issuance reduction chat but it sounds like people are trending toward an 8-month Constantinople Forks so if you go to the link for the Constantinople progress link and the agenda that actually defines all of the EP that are gonna be going into

Constantinople and no other ones are gonna be going in okay great so let's move other than potentially the issue introduction yes mam yeah exactly okay yes potentially those so that would that would be something we would need to decide on today because that would also affect testing am I right Dimitri if we do an issuance reduction that's gonna affect testing pretty greatly I mean if you disable some issues from Constantinople to being included into custom yes yeah let's let's just to color and see totally okay so let's get

into the chat about potential issuance reduction so what's going on is there are multiple competing a IPS to delay the difficulty BOM and then reduce or maintain the difficulty are the black block reward 858 is to delay the bomb all of them are to delay the bomb so that's not up for discussion that's already been decided last meeting however some of the things that are brought up are to reduce the block reward to one F or eath per block that's 858 one two three four is to reduce it to two eath per block and then twelve ninety-five is to keep the reward at three eath but to change other

factors such as the powa incentive structure involving uncles and nephews so if it especially for 1295 that one goes a little bit more in depth because it is not a reduction if you all could open that and kind of browse over the motivation and the abstract because that'll be a more technical argument but we'll start with just a quick introduction of the people I've brought on to to talk about this stuff so we'll go alphabetical order I guess and just talk just twenty second introduction about who you are and who do you feel you represent and if you support one he IP over another so basically there's people here who are

miners there's people here who are investors and there's people here who are from the F minor or Prague POW community so we'll start with Andre yep I can hear you great yeah late developer from the earth team a minor team and I'm just digging into the EIP 1218 sorry in the 1295 which I found it very

interesting from the miners point of view I think that anyway we should speak before those AIP is which may be the economic implications about the reduction in the reward or keeping of it actually given the conditions of the market I don't believe that a reduction of the reward would help anyway the miners whether they use an algorithm or another or they they they use any AZ

compliant devices or whatever this is my personal opinion I think I will anyway here other opinions before adding some some something else sorry for my English but it's very good it's very I have very tired yeah thank you no problem Brian you're next good morning everyone my name is Brian ventura i am the CTO of atlantic crypto we are a large GPU mining organization based in the u.s. prior to founding atlantic crypto I ran a systematic fundamental natural gas and electricity hedge fund based in New York prior to that I worked for a large environmental

asset manager where we invested in Kyoto Protocol Clean Development Mechanism projects throughout the developing world I believe net net we reduced carbon emissions in the atmosphere by over a hundred million tonnes I have a deep expertise and energy and nascent commodity markets not only do I feel like I'm representing the mining community here I feel like I'm in I'm representing the sophisticated investor community I clearly support the IP 1295 it is re IP I'm happy to go in in the in depth discussion later on thank you Brian Alex thorn for I guess from fidelity labs alright Alex might not have his mic were

his or her mic working Carl you can go next yeah hi hi my name is Carl arson thanks for inviting me on Hudson and so I've been involved in the etherium community since the beginning I started the earth Draper subreddit and help moderate that and prior to that I was a well I have been a web developer and prior to that I was building wind turbines in Afghanistan so I authored AIP 8 5 8 back in January and the the concern the justification I had used in the ERP is really to look at the

environmental impact that that mining has on behalf of the network and and and you know seeing that we we should do what we can to reduce or mitigate that that environmental impact Thanks alright thank you so much next we have Eric Connor an investor and part of the community since day one so I am here representing a IP one two three four or I guess I should say supporting which is a block reduction to two as well as the way in the difficulty bomb because I

think currently the network is grossly overpaying for security so again thanks for having me instant wonderful all right so next I'm not sure who hacked our is hacked are could you introduce yourself into a flare oh okay no problem same with Jason temple I'm just an observer here thank you though okay

perfect Mariusz are you an observer okay next we have Matthew light am i coming through yep you're coming through great okay I'm Matthew light a software developer a longtime interest in block chains and aetherium in particular I did the initial VIP for the initial issuance reduction from 5/8 to 3/8 and now I'm in support of an additional issuance

reduction the reduction from 3 to 2 I think feels like the most prudent response to the situation and that's the IP 1 2 3 4 by F riche should so I'm in support and a number of other investors that I'm in communication are all in support of 1 2 3 4 issuance reduction to two East per block okay thanks so much Matthew and what we got next is Tim Coulter yeah Tim Coulter founder of truffle

I represent small to medium minors so I've been mining started out in my garage when if your am started but now I have about a dozen machines looking at a 100 about 100 GPUs I also have a bit main miner I bought that completely as an experiment so I have first-hand experience if you guys want to know about I'm in favor of a block reward reduction I don't have a preference to which one most I am most likely going to continue mining either way purely out of speculation I think depending on price

fluctuation I might decide to take that money from electricity turn off the miners and just buy at the market depending on the way things go but beyond that you know I have no nothing is likely going to stop me from mining one other piece of contacts I'm up here in Washington state electricity is cheap most most of our electricity comes from renewable renewable energy okay thanks Tim and then we have Chen Zhen CEO of sparkle which is one of the largest mining pool in the world country hmm we we are supporting more than 20% of the

heart rate currently avi theorem with with spark for brand we we control now we control or we have about 300k rigs involved and connected to our money for every day um my my personal experience is aqua has been 2 years and before that Jasper port is actually ordinary from East fans which is actually the the first and the largest Chinese if they're in Cranford community so we did not start with a money pot we start the business as a side project as we're beginning after all of course I support with the delay of the Ice Age typically

pumped and we prefer VIP one two nine five if I have to choose one two three yes okay thank you and then I have a quote or a statement I suppose from ether mine Peter from ether mine he they control 30% of the hash power and they said the following we have reached out to our miners and from their response that is clear that the most important point for them is to include a Pio Pio W change to obsoletes Asics regarding the block

reward their opinion is somewhat ambivalent some prefer an increase obviously some argue for it to remain the same and a small fraction want it to be reduced regarding the difficulty bomb most of the argument for a delay most argue for a delay and not complete removal this is also the position of ether mine the position ether mine will take as a proxy for our miners Hudson we also have Alex thorne here posted a text looks like as Mike's not working do you see that in the chat stream here let me look for it it's not working we're ready to take a fir we aren't ready to take a firm stand on these E IPs but in general

we favor 1295 ok so thanks everybody for the quick intros we only have about 35 minutes so we're gonna have to get next on this oh yeah I skip you yeah you did sorry go ahead no problem my name is John by the way John seer Heath miner dev full disclosure I'm neither a holder nor a miner I'm purely interested in the technology and the development of that technology and participating in in development in general you know my test rig generates about 56 cents in profit a day so clearly my motivation is not

financial so you know given about ten hours to prepare for this eight of those being spent sleeping I don't have any much factual support for any arguments might propose but generally I favor the third choice which I believed was what 1 to 95 a slight reduction on uncles and leaving the rest as it is my basic my basic feeling about the whole thing is that the long-term viability of eath is basically predicated on success of you know scaling that's really the bottom

line and this will be the only thing that will D couple eat from that other big coin yeah that's about it Thanks thank you very much you also missed [Music] I asked about Marius and the mic wasn't working is it working no problem Marius is your mic working ok if it turns out to be working well we'll get them back in so there's a good number of people who've been talking about 1295 so I want to get just a really quick couple

of minute overview of what 1295 is just pretty much just the abstract and any part of the motivation that you want to elaborate on I believe this is Brian's is that right that's right so yeah he 1295 the the bottom line result under status cooperation of the network is an eleven point four percent issuance reduction it addresses the issuance reduction through incentive through a current incentive misalignment where the mining community whether they're doing it or not is currently a sent device to maximize uncle rates the greater number of uncles the network has the higher the total issue and it's tired the total

aetherium and u.s. dollar award currently fourteen point four percent of all aetherium issuance goes to ancillary rewards like uncles and Oakland inclusion reports ew 1295 adjust the uncle structure to incentivize up will note configuration and decentralized mining when I say decentralized mining what I really mean is that anytime you have a large pool you're centralizing the delivery of work packages to your miners anytime you do that you're going to increase latency to your miners it's just the round-trip time it takes to get that work package and to deliver a share that will increase Uncle rates ew 1295 proposes to reduce uncle rewards by 75% and to only

reward the first two levels we believe this is appropriate because you shouldn't be rewarding rewarding subpar infrastructure it targets a sub six percent inflation rate by q4 2019 it leaves top-line incentives which is the block reward in place to minimize security risks the network as we go forward going forward we have a couple of points here we'd like to make one we suggest that we adopt a I tweet yeah be 1295 now we believe it's in line with Danny Ryan's Casper FFG VIP proposal we're and it's pretty close to the issuance schedule therefore year one we also believe it potentially provides

natural natural issuance reduction when the nodes are actually going to be optimized to reduce the uncle rate we think that rate can get down to 5% five to 10 percent which is pre metropolis we think there are a couple extra steps after this the second is to commit to a change in the proof-of-work algorithm which I believe we're going to talk about later I think we target a q1 2019 roll out for that and the third and most important piece here that we'd like to talk about is after this is done and targeting the heart for work at this time next year we believe that community really needs to come together to commit to a monetary policy that has a set roadmap for issuance reduction and should be adopted we have initially

wrote our Goedecke to diffuse the difficulty bump we'd like to change that position to just delay it so that that is a forced conclusion next year and that's kind of where we stand so happy to field any questions or anything you guys does anyone have a comment on that and a particularly technical or more technical slash economic crypto economic comment on what that would do to the network regarding a long term policy before we do that can we get a high-level overview of all of these before we go into any depth on any of them ooh that's a good idea the author of 858 can you give a quick overview of yours is it just reducing it to one eath and what's the motivation

yes and actually I've made a pull request recently to also include delay of the difficulty balm just into twelve months so that's pending a pull request on the master repo and yes if a 5/8 is to reduce the block reward to 1/8 per block it was written in in January and the rationale is is just looking at the the costs on the network in particular the environmental costs of the network from mining and and wishing to to reduce those and looking at the reward is the

lever to to do that okay thank you and then offeree if you want to go over one two three four quite similar to eight five eight I proposed to reduce the block devoted to either and to delay the difficulty bump by three million blocks so around 1.5 years it's a bit more moderate as my rationale was basically to try to keep the current proof-of-work assurance as possible okay thank you now if we can go back to Danny's comment

because I was interested in his perspective on it oh there was a mention that we need to have a long term monetary policy that's really hard to do without having transitioned to prove mistake yet because prove a snake radically changes the amount with which we can pay validators in the network compared to the miners and so from my perspective a lot of these transition points between now and and then our kind of stopgap GAAP measures to keep the network healthy to keep people using the network to people people posit about the network

keep the network secure until we get to that point and so I don't think we can begin to really talk about long term monitory colony until we're talking about person to take and Sheridan hey Danny it's Brian so I understand those points we've got some ideas that we're gonna roll out after we get through this fork here of ways to kind of I'm gonna say introduce monetary policy to deal with the stopgaps our main goal here is we don't want to have an unlimited amount of issuance before proofs take happens right as we believe there should be a hard cap of issuance that goes to the mining community it's up for debate it's gonna be a long debate it's kind of

why I set a one-year time frame for that to be rolled out but I do think that is a very important conversation and I think that it would be very beneficial to the whole community if we didn't have to do would do this every three months or six months and just kind of had that roadmap everyone agreed - we can talk offline about it I think this is I just hope this is the start of that conversation I just want quickly add one thing to this so I heard the same comment from some of the F minor deaths that they would much more prefer to have a hard cap of the supply down manipulating day the block and anyone here already on Bitcoin actually reduced got like it

burn low the riddle of please comes very did you speak up a little bit neck it becomes very cost-effective for miners to mine ultimate chains and the same results I believe will apply to etherium we can't there's research that shows we can't reduce reward below a certain level and expected stop okay do we have other comment on any of the EIP so far can I just make a comment about the IP 85 a please so hey my my career started as an environmental investor and the group of partners that I have in Atlantic crypto

what they did the first emission reduction transactions in the world in the early 90s I believe that the reduction in emissions from reducing from three to one would be on an absolute basis probably zero the hardware is going to run it's going to continue running on other networks the incentives to built a new hardware may not be there but those incentives aren't there right now so when we look at the emission kind of impact of the etherium network it's about 4.2 million tons a year the etherium network can be powered by one combined cycle natural gas plant in America and you look at all the large

miners everyone wants to find renewable energy resources as fast as they possibly can because it is the cheapest most available power the the environmental implications of the current network versus Bitcoin are I would almost argue de minimis just wanted to make that point okay thank you let's go to someone else oh go ahead Peter I just wanted to nitpick a bit with this argument I don't think it's it's a good strategy to say that class since Bitcoin does it we can also do it a little bit Cohen does it worse so nobody cares about ether so I if we can for example reduce emissions or

not I don't think we necessarily need to to just do it because I do analysis doing it so that's I don't really accept that line of thinking okay yeah oh go ahead Carl so I my mic was muted I just wanted to respond to a few of the things that that we just talked about in terms of the minors leaving and going to other networks and I would ask to look at you know by my calculation aetherium is roughly 65 percent of the of the GPU mining I mean it's sort of the market cap of aetherium dwarfs the combined market cap of other GPU my noble coins

so you know maybe some you know some miners would go to other chant to other coins but but you would expect that if the demand that that the etherium network is is sort of asking for this money if if we remove that then it's going to remove it from the wider pool the sort of total pool of miners and so I would expect some mining definitely expect some mining to to drop off totally so and and the the electrical that you the electrical consumption that

you can attribute to that mining you would also expect to be removed and and to the point regarding using renewable energies I I don't buy that renewable energy is always used for for you know for to power miners and even if it was the case then actually the miners are consuming electrical power that could be diverted to other uses which of now probably being a burning fossil force to provide fossil fuels to provide the electricity for so thanks arguing about the origin of the energy

used to mine and I don't know if we should argue about where the Energy's going is that kind of your point Nick pretty much I mean if if we can reduce the minor and seems that without making security we should do so regardless whether it's good or bad influence and likewise that become friendly yeah so that's kind of the approach I'd sent I think my blog post got put in the chat here a few minutes ago essentially I was taking a look at the IP one two three four which is a reduction from three to two I took an approach where I thought the participant participants of certain networks should be willing to pay for

security based on the value of that network so I took that point as a kind of a bench line or a benchmark for that first ethereal level set the market caps and then looked at what the network went aetherium it's currently hanging out in minor rewards and if you look at that appearance currently hanging out a factor of two and a half times higher and so he dropped that block reward down from three to two those actually start to come to mind so that was kind of my argument for one two three fours more about hey we dropped one Greg kitchen me please I've run I just want to say - seems like a reasonable

compromise to me and an investigation and maybe a somewhat of a commitment to try to do prog pound eight months also seems like a reasonable element I go take care everyone bye-bye Danny listen just to respond to Danny's comment there I think that going from three to two is going to do a couple things to hardware composition on the network the the community seems to have presented a pretty strong mandate that it would like the proof-of-work algorithm to be shifted from the current ASIC I'm gonna call it friendly to a less friendly algorithm like programmatic group work reducing from three to two in the top in as that happens is basically going to

force the GPU miners off of the network and the only thing that's going to be kind of left behind is the ASIC community I don't think that the community wants to do that overtime here before that happens a lot of these things have already been produced and purchased but if our goal is to remove the ASIC participation from the network let's not do the opposite I'm all for an issuance reduction when we understand what the ASIC implications are the ASIC participation the network was which may be the middle of next year the end of next year as part of my idea behind the monetary policy roadmap but I think that right now

without putting security the network at risk and the composition of the network at risk reducing to two is probably the wrong move I believe that 1295 makes a technological advance in changing the incentive structure for miners and mining pools while also reducing issuance but by a pretty sizable amount if you go to from 3 to 2 and you don't change uncle rewards you go from 3 to 2 the network is going to be incentivized to have the highest uncle rate possible right issuance will be higher than what we can ask a question sure class ask the question first real quick and yeah just

in case somebody who can talk a bit more to the to the Asics that kind of running around I mean my impression was that the ant miner III the running efficiency is about on par with current GPUs and the a-10 is that is not is not even released released yet so do people have data on the a-10 because it it claims that it's gonna have you know sort of a 2x improved running efficiency you have read yeah let's have Tim go first Tim what was your perspective I don't have an obviously donate 10 or have any date on the a-10 I do have an e3 what I've noticed is so my GP rigs are all a gpus each

rough cost in given all the pieces that we have to put together is about $2,500 the price for the III when I bought it was in the first round which was 800 bucks but still around 2000 so there's a there's a smaller a cheaper cost the efficiency after you tweak the the GP rigs to work the way that they should the efficiency is roughly about the same especially if you get the right GPUs that you can overclock so you're getting about the same hash rate out of it now the thing that for me the only

difference like we reduce it to two or even one like maybe I'm a special case of mine or something I'm gonna continue mining with use TPU rigs however the thing that the III machine has over a GPU rig is that the GP rig has you know 47 different moving parts and so management of those that hardware is much harder than just plugging in the III machine and letting it go so I don't I don't know the art for me the argument whether or not you're gonna shut the GPU [Music] miners out I'm not sure I believe that argument however the only benefit that the III one has is mostly management how

much time it takes from thank you guys this is Brian again so the the early information on the a-10 is that the the top-of-the-line model there is gonna be able to do 453 mega hash at 850 watts which is about let's call it three times as efficient as current as a GPU rig we've got a couple GPU rigs we've got probably three thousand of them so before machine learning artificial intelligence and those use cases are just beginning to show up and they are projects on the etherium network there are folks who are looking at this hardware and saying that's a resource I can use he a network of Asics

if when this network does go to proof of steak very not usable right so incentivizing are allowing the build of a network of Asics I mean it seems to me like you're leaving nothing except for waste behind we we made the choice when we started this business that we did not want to invest in a sick mining in aetherium and look at ASIC monitoring and aetherium because of the flexibility of the GPU we still believe that I just thought it was a good point - sure can I wanted to can I make a further point um when I go ahead and then go ahead come back I was gonna have Chen give a hit their perspective from the spark pool

perspective on some of these issues if you have any sure there are two major points I want to say and it says China angle firstly is about environmental I think it's to ideal to think that that's the GPU miners out or cuts to the heart rate will help with environmental problem because at least from my understanding all the Chinese miners are using non-renewable energy however those energy are all wasted otherwise because all the energy are produced no matter what it's not because of the mining it's because of the electricity Trent is over there it will never stop however the Chinese miner

happened to use the mining business to use the wasted energy so which I don't think cutting the issuance will help with the environmental part and it's not in China second I want to point out is it's more about is from my personal perspective but it's mobile network security level I post a link before in the chat about crypto p51 app which is a website how q to calculate how much it costs to actually do a to actually do a 51 attack to the two network theorem it's not that much however while there's no attack ting if their network I think

from my perspective or study as a 14 years old there isn't any uh shred you can use this market to attack experiment that work so like there are a lot of maths people doing on the internet about what if we cut stations how much heart rate about roll out however and is to me what it's not like if you are going to cut three to you and only only one-third of ahartry will go out or maybe well well cuts if your network security to two-third it's more nicer it's Monica now we have hundred soldiers and that's we have in the world but if we cut 40 of

them out then it's not about if we have enough if the enemy or the fat people have enough soldier to attack this because before they have zero now we have forty they have the chance they have the once they have the capital we can always have the choice to hire those soldier all those heart rate to attack them so my major point is if there never has been secure or safe to to now it's not because of attacked a theorem cost too much it's simply because of not all heart rates are on if you're in network and no one has extra heart rate to attacked it and even though we have enough money it's not gonna happen that's ninety major points so far okay so some some of the things I'm hearing

are actually Karl you still had a point to go first yeah well the point I wanted to make was that going forward we're we're I think we can sort of concern ourselves with the the current hash rate and the current miners and their running efficiency and to you know in order to you know we don't need additional do we need additional hash rate like are we talking about kind of wanting to grow it continue to grow the hash rate is that is that desirable in my mind that's absolutely not desirable and and and and in that case looking at the kind of the the equipment cost and actually we want

to be discouraging that equipment that investment in new equipment and and one way to do that is to is to you know lower the lower the the block reward and that that you know so worsens the economic argument for that investment can I make a quick point on that I don't understand the rationale for building either a sick or GPU mining rigs at this point my calculation on the III is that it would take roughly two years to pay for itself unless something extraordinary happens to the valuation am I am I getting this right for me

[Music] speculative at this point yeah it's kind of the price to an extent and I suppose you're right yeah but even at say 5 X to the price it's still hard to make a an argument for it but anyway just just my thoughts rather study from very specific or comprehensive article about why a sick miners exists and how the ASIC

manufacturers making money essentially on the ASIC manufactures they don't they are not responsible to the miner once once the miner boxer rigs out their profit and and the costs will be paid off right after people made the pre pre bio whatever so essentially all the money well be correct all the capital and other prophets before the miners star money well starts before the ASIC star mining so of course they have the have the incentives to manufacture plus some the future of the mining does not really affect to them because once

it's produced once it's designed manufactured a lot and we're not talking about the manufacture incentives we're talking about minor incentives the people that actually purchased that gear why would they there's been no incentive for me to continue to invest all this a while ago with the full knowledge that mistake is going to come and I'm ready to shut it down so if people continuing to invest in mining I completely agree that we should be incentivize them okay I'd like to I'd like to talk a little

bit about okay I'd like to talk a little bit about sort of our background of why we think a reduction is appropriate we did a reduction from five to three I think probably most everyone involved with etherium would agree that that reduced issuance so we'd reduce the amount we were paying for security and we did not have a negative effect I think a moderate reduction from three to two would would have the same effect not impact security significantly but we would reduce the billions of dollars a year that we're paying for mining proof-of-work mining which is not where

a theorem really wants to be anyway we've had a significant setbacks and delays and moving the proof of stake I was looking at a thread from mid 2017 or Vitalik said he expected a the likelihood was that we would have proof of state by the end of 2017 now we're talking that we're hoping it's going to be in 2019 I'm certainly hoping that as well but getting the issuance under control to a level that's comparable with with the other major block chains in the space will have good effects for price which is important for paying

developer salaries and projects funding new projects it'll also reduce at the margin that will reduce the incentive to put more buy more mining equipment more or hash power more electricity we're currently way over paying the miners based on you know we're basically talking about seven percent seven and a half percent per year of the market cap of the cerium is being paid to miners that's that's too much bitcoin is paying a lot less this is a competitive marketplace for crypto in a

bear market like we're in now a lot of that especially from the big miners from bit Maine etc almost all that is going straight to the exchanges and getting dumped and that has a negative effect on everyone's ability to keep working in this space over a long period of time so we don't want to be radically overpaying for something that we're trying to get away from anyway yeah so we don't have a lot of time left so what I wanted to do was this I from talking to certain amounts of miners I've seen that there is a greater need in their opinion for a switch to a different proof of work

algorithm in order to make Asics less profitable or not profitable at all so miners in the room one by one if you cough hash that would make mining with Asics less profitable or not profitable at all so basically could there be a compromise where there is an issuance reduction but also a switch to something like prog POW in the next eight months or ten months actually is that something that would be agreeable we'll start with Brian what what would we think on that so I'm in favor of an issuance reduction paired with a change to the program to the proof-of-work algorithm the caveat with that is that I believe that 1295 is a

great stopgap before that proof-of-work change happens if the community would like to go from make an additional reduction at that time I'm in favor of an additional reduction when the proof-of-work change happens we just don't understand the extent of the ASIC participation at the moment so I think that a good compromise in my view would be 1295 today proof-of-work change at some later date along with an incremental reduction on the block okay and I have to point out so in general other talks we had today is nice to have

but now our security is not being here we cannot just use single simple math to calculate there is a tipping point once we reach that tipping point everything will break down and instantly and when it happens we can never move back to the talking that's that's one thing I really want to make no matter what yeah so it sounds like our Asics a bigger threat or as an issue at issue its reduction a bigger threat to that security in your opinion in my okay the the the trend officer as well be a big big impact to the security because no I know no one can simply calculate what's gonna happen in the future with math okay well we can

we can look historically we can look historically it's what's happened in 2017 you know we dropped from five to three and you know you Asian was largely what it is now really you know the the the you know that was January when when things you know showed up beyond kind of the 300 level I have a I can post a a link to a chart that has both the and there has both the hash rate and the the issuance but the issuance converted into the eath in dollar terms did you have a comment yeah was more in 2017

there was at least my knowledge not that widely deployed ASIC usage so reducing the issuance reward really just kind of wasn't even even playing field for all whereas you know doing it now only the strongest ie most profitable miners ie Asics would survive but Asics don't have at the moment nobody's demonstrated that a zzyx have a running cost a significant running cost in fact any running cost efficiency improvement over GPU so human that there's that they don't you know yeah from what I've heard Carl is that they're on par with more the top of the

line rigs GPU rigs that are out there now but for those of us who have legacy you know systems going back you know a year plus they're they're significantly more efficient and then and you have to think about the future Asics that are gonna be much more efficient the ones that like we would be breaking in about ten months I have to question that assertion because I mean the ethos or the remains memory bounden and there's no magic in ASIC I mean I've worked for most of my career in a semiconductor company the the real issue with E is the memory bandwidth issue and it all goes with the memory technology not the the underlying ASIC I mean GPUs are Asics

once you get down to it yeah so it was prog pal have you explored prog pal and is that something that would be a solution or not not really a price well it does a couple of things I think it reduces a dag size and it tries to exploit more of the circuitry that's that exists on current GPUs of you know making it more difficult for an ASIC to reproduce nevertheless possible and also if you think about it the shortest time you can deliver an ASIC is about three months it's a development cycle of three months cycle with the fab that's not a long time I don't think it would take the asic guys much time to catch up with that and secondly I'm concerned with the

introduction of a new and you algorithm in that what I've seen bug pile is that it's unoptimized at this point so we're basically getting into this kind of you know arms race with the minors see who can out hash who and you know various private and public implementations I'm very much in favor of open-source you know it's the great equalizer so I tend to favor leaving things as they are yeah and also someone skipped a little insofar as I mean I'm not looking at the list of randomly selected operations it

says many of these operations which don't require many days a couple of them of shifts which actually require gree dick yeah I'm also skeptical yeah currently sure I think Mayans mentioned that for example changing the issuance would change the hardware distribution of hardware build up of the network's button and arguing that sneaks probably me to take closer look at proof of work

from the same perspective because currently so miners don't just blindly by all the GPUs they can get they probably benchmark them and probably selected those that had the highest memory bandwidth and reach the highest hash rates now at when you are actually changing the proof of work algorithm it it means that certain GPUs that were insanely powerful beforehand might be insanely suboptimal because maybe they have an insane memory bandwidth but they have a crappy computational unit and basically my point is that it kind of seems it seems a bit naive to say that if we to switch from et hash to pro proof-of-work that all of a sudden miners with GPUs will be happy because

I'm quite certain some of them will not get happy from the results so that's probably something that would also need to be investigated is now yeah I think we should also we should also look into the new developments with like this gtx 2080 and stuff like this this is just coming out so maybe we should wait for benchmarks on this so that we can evaluate the changes that we make before we start making them interesting ok so we're pretty much out of time we'll probably go on a few more minutes but what I was wanting to get some suggestions on from everybody here do

you think does anyone think there's an optimal way to figure out sentiment and to make a decision on this the eternal question yeah exactly if anyone has any ideas throw them out there I didn't really think anyone would have a great ever Oh Greg a useful straw poll is yes no and over my dead body so you're talking about rough consensus there we you can handle no is over my dead bodies you've got a try to avoid yeah who's who of the mining pools or miners here would leave if we did an issuance reduction no

no not I so if you uh it depends on the reduction cuts them okay if you go to one we're probably I mean we've got really low power prices and we're probably not economic at one yeah at two we're going to explore every alternative use case we possibly can without a shift and proof work yeah cuz I not I vote was in line with what Brian just say going from three to two so and would this be well this also be dependent on the price of ether though absolutely right we're just talking about this moment in time I think that the part of the reason why we laid out our our path to go kind

of 1295 today and then an additional reduction if we change the proof of work algorithm as is the stopgap for that right it minimizes the security risks during that time and it still gets everybody kind of where it seems like consensus is going in the medium term that's our opinion right Andrei point I didn't catch from the other participants the very simple question is is it worth the effort I mean is proof of the state coming okay it's a mix of proof of stake the balance of work but if we leave the

basics to enter freely in this in the security of the network and we let the small miners get out due to costs are due to increased or decreased efficiency of their GPUs we will have a network we will be which will be in the hands of the Asics not a problem per se but it will be very difficult to keep or change in the next steps the proof of work algorithm because if you do that all the Asics will be unavailable for a change in every in every aspect of the

algorithm so you might end with a very secure network I think it's very debatable to say that miners will drop off with a reduction expression at you I mean any any studies that looked at inflation 7.5 percent were overpaying for network security by factor 2 and a half to Bitcoin you know we should it was generally consensus that we'd end up around 100 million total supply and about 1% from there or the one in hat a tripled since the last reduction so I just I'm not okay Chang do you have an opinion on where

the breaking point would be as far as spark pools mining operation would go if you all have what however currently I think the top I post has more than 70 percent something like that however once the issuance changed I don't think the post post you have that much power or

weight to support anymore because once the issuance change the - won't leave the pool to do something else but once those hot power goes to market anyone who is enough capital I'll have the money to purchase the power and the tax the network it's important that so still my point is this the network security power or the problem is not not on me no you cannot just cover it with math understood do we have any other final comments I honestly don't know how to make a decision on this so we don't really know where we'll go from here I have some maybe we should just wait and

evaluate Tom this this new proc proof-of-work works and maybe look into this a bit more and then come back and then make a decision yeah so the problem with that is that we're kind of so we kind of need to make a decision for Constantinople hard fork and yeah we're on a time crunch there so this isn't something that can be decided I don't even know if we can wait two weeks on this what what do you think Dimitri up he's gone never mind

yeah what do you think you say Martin yeah I would like to propose that as we draw near to commit to an action hard work that we have more regular meetings and perhaps have that every other week we have a technical meeting and discuss the you know test cases implementation where we ask potential details that need to hammer it out in the eaves and then every other week we have I don't know more general talks like we're having now high-level talks okay so you're saying that we should have a meeting in a week yes I guess what I'm saying

I'd be down for that it's in the technical detail and fix all those quirks that we don't really have time to talk about on these calls I I would second that that gives us a week now that we've heard everybody's cases and motivations to dig into the eats read to begin in that context and then we can talk about it again next Friday and make a make a call okay I think that's a good idea there isn't really any other ideas out there so I think that's what we're gonna end up doing so we'll have another meeting in a week of the core devs and we may or may not have some of the

people some of the guests from today on what we likely won't just because it's gonna be a rehashing of the same arguments so between now and next Friday there's going to be a lot more community comment on places like Reddit and Twitter and the fellowship of a theory and magicians forum probably about the e ip's and about the different perspectives here I want to thank everybody for coming in and giving their perspective I think it's great to hear all these different voices from different stakeholders and I think we got some stuff accomplished today just hearing some some in-person conversation over the different viewpoints so we will see the core devs in a week for a meeting to discuss some of the e IPS and

more detail and hammer out some details and thanks everybody for attending we'll talk to you all a week thanks everybody [Music] [Music]

[Music] [Music]