ETH 2.0 AMA with Danny Ryan

Transcript

the banquet state the nations are brought to you by wiring wiring is defy's first self-building community run project which i just get really really excited about iron is a system that seeks out yield in d5 and it does that in a number of different ways a very aggressive way is with the vaults where you can deposit your preferred asset of choice and different defy experts will come in and generate a strategy for what to do with your deposited token right and so it'll go find ways to get yield in that deposited token in d5 for those who want to just earn yield on their stable coins the earn system is for you where you can deposit your preferred stablecoin and wiring will go and figure out which money market on defy and defy is

producing the best interest rate whether it's dydx it's compound or ave it looks around d5 to see where the yield is coming from and it directs stable coins automatically so you don't have to check them out at yearn.finance to get started and also check out the stats page to see what other people are doing as well we're also brought to you by monolith monolith is your cool new defy account your defy savings account your default checking account except the cool thing about the monolith defy account is that it gets software updates right you actually get to increase the usefulness of this over time so here are some of the features monolith is a smart contract wallet with a lot of the features that you would expect if you've come to know defy and what it is you can you can add money to

it you can put that money to work in compound and accessing yield uh but you can and you can also swap through uniswap what was cool with monolith is that they will send you a very sexy monolith visa card that connects to your smart monolith smart contract wallet on ethereum so it's a really awesome tool to live a bankless life with a a savings account that gets software updates so this is something that you're never going to find out in the real world but you can still do real world things with you know real money and like buy your groceries so that's just fantastic coming soon to monolith actually already here to monolith is now you can buy die and get it sent to your wallet directly right so it's also being an on-ramp so you don't

have to go through your centralized exchange like coinbase or gemini or wherever you can just go straight from your bank account right into your monolith checking account smart contract wallet so check them out at monolith.xyz all right welcome bankless nation to another ama this isn't ask me anything we've got danny ryan here whom we're going to introduce in a second quick logistics here's how this works you can view live on youtube it's broadcasting now you can ask your questions on youtube if you're a bankless member you get prioritized questions in the bankless discord so make sure you start pumping your questions for danny we're

going to be talking about eth2 we're going to try to wrap this up in about an hour and i will interrupt at various points and let you know how much time remaining we have this is a super important ask me anything because of the timing eth2 is coming it really is despite what they say on twitter we just had vitalik on the podcast to talk about the why of aetherium now we want to do this ask me anything with danny ryan to talk about the what the how maybe even the when we'll see without further ado i want to introduce you to danny ryan who is an eth2 researcher a coordinator of clients a herder of cats he is the guy to ask about eth2 progress danny how you doing man

great thanks for having me excited to be here okay let's start here uh you know we get dipped on the first question on a scale of one to ten how excited are you about eth2 right now one to ten um i mean definitely ten um i've been working on this for years um it i can't believe that is true but um even before e2 kind of took its current form which it's been in that form for over a couple years i was working on um proof-of-stake stuff uh in a different form and so um i am anxious and excited to get this thing out i actually i want to i want to tell you a little a little story so in japan they have these things called darumas and these little dolls and you draw in

one eye and you essentially you make a wish or you make like a goal and then when you're completed you you color in the other eye and it's like a it's a blind buddha that like achieves enlightenment or something um and you so i have this little white one little white daruma and i colored in one eye in the beginning of 2018 because i was like i'm going to ship casper like we're going to ship it this year it's the white daruma it's like a little ghost um and this thing it usually sits right there but it's not there right now um it it haunts me right it sits there and it looks me and it has one eye and it had to shift it when we shifted to e2 and sharding and all everything combined um and so i'm i'm very excited to take this little daruma

and give it its final little eye and then you burn it so that's gonna be fun too fully formed and that will achieve enlightenment is what you're saying is that for yourself or is that for all aetherians yeah anyone can share in the enlightenment i i will gladly share danny you had a really fantastic talk at the online summit uh not too long ago where you talked about how the whole eth one into e2 transition migration is a misnomer and i thought that was a really fantastic perspective that i i wish more people had so i kind of want to give you the opportunity to kind of uh pitch to us why why is it not a transition why is it not a migration and what is it really and and how should we really think about this thing

right so check out this talk um it gets a little it's like not incredibly technical but like gets into more of the structures of these things um but we talk about this thing called eat one we talk about this thing called eat two all the time um and what we think of as eath one is like ethereum as is today and what we think of as e2 is like this thing that we're trying to build which has proof of stake and sharding and scalability all this fun stuff um and i claim that um we made a mistake in the naming and that it's a misnomer and it implies this like sequentiality of the systems meaning like you know we're going from windows 98 to windows 2000 it's like this whole new thing or we're going from um like a car or a bike to a car or a car

to an airplane where is it really it's it's a more um it's more we're talking about layers of the of the system um so what when we think about eth1 uh all of the like complexity like you have this little this tiny little proof of work thing it's like from a software perspective relatively simple all the complexity of each one all of the hard work that's gone in for the past five years to iterate and upgrade and make this thing like what it is is is in this like what i call the user layer it's in execution it's in transaction management it's in state uh it's in state sync and all these things that like make the user layer of ethereum awesome and and also a lot of things that make ethereum hard where the proof of work modules just kind of stayed stable what eth2 is is really

taking this like tiny little proof of work thing that's like the brain of the system and saying can we reimagine can we redesign the brain of the system to be more secure uh more decentralized and more scalable and that's what you know when we're talking about the e2 project it's really we want to replace the brains of the system uh with something more sophisticated rather than replacing like the guts of the system like it's still the same system um and so what we consider eth1 that state that that user layer um is to exist kind of with this new brain what we call e2 um and so there's probably some better terms to use there uh maybe we'll just call it serenity the upgrade of ethereum's core consensus um as it probably should have been called the entire time but

uh so that's that's a little bit of that check out the talk um it's something that's important to me because uh it makes it hard to communicate about these things if we can't talk about them um and so check it out maybe we can all begin to talk about the separation of layers in a more organized way so just to kind of summarize that and make sure i'm on i'm on the same page uh with ethereum there is the stuff that we do right the the defy yield farming unit swap any anything that's a contract right tokens transfers that's the stuff that we do and then there's the uh infrastructure that supports that right and so the stuff that we do isn't changing that's going that's

coming along with almost no changes it's the how how the machine of ethereum comes to consensus around the stuff that we do that is uh being like swapped out and so the way i've explained this to a friend of mine is kind of like the the indiana jones uh the like taking the idol and the the sandbag and swapping the sandbag on for the golden idol and then at the same time is that an accurate like illustration or maybe is that did that lose some nuance that's probably reasonable um i'll give i'll give you two others that like i think are kind of reasonable as well um joe lubin actually said this in a talk i don't there's been like this this whole covid online thing is like all blur so i have no idea what it was or when it was but joe lubin um was

uh in in some sort of like online thing said it's it's like and i actually i think mistakenly used this analogy earlier but um it's like a modern operating system your upgrade your modern operating system what happens usually if you choose to like it feels very similar like you brought your files you brought your programs you bought everything and like it generally works but maybe you have like some upgraded hardware maybe you have some better security features and maybe you can like if you're a developer you can like get in and do some like new cool things and maybe you have a bigger hard drive but you're still working on what seems like the same computer your keyboard might click louder because you're into that um but you know so that's that's one way you can think about it is like

the user layer your user experience can be very similar but if you choose to can be enhanced over time um and then another one is and i don't love this but it's because it's like weird like war analogy but imagine like ethereum is like some fighter pilot plane thing jet a jet uh flying around it's doing awesome stuff um and then like what we're doing is like essentially preparing the aircraft carrier and it lands and like now we have like a bunch of other jets that are like all managed under the same system both of those kind of suck but you know different ways to think about it i'm still i'm still working on the memes and the analogies but maybe [Laughter] yeah well fantastic yeah and we'll add

the indiana jones one to the roster there david nice one there um all right so can we talk up for a minute because i think this is a huge question in the ethereum and in the bankless community about the state of the test nets so for people who have not caught up right i guess maybe zooming out what we're talking about then the next thing that's coming is uh ev2 or serenity um phase zero right and then we've had this series of test nets to get to a mainnet phase zero can you talk about i guess phase zero and the the feature set of it and then the state of the most recent test nets that have happened and kind of where we are

now right so what we're doing right now is we're bootstrapping the new consensus for ethereum we're bootstrapping this proof of consensus and that that's called phase zero um to be honest like i think these the terms these phases are like nice for technical discussions but i don't think like are incredibly meaningful for the community so i'm even just trying to talk about them differently um so what we're doing at the end of this year is we are bootstrapping the new consensus mechanism uh ethereum i don't is like a 50 billion blockchain i don't know depends on the day depends on the hour depends on the minute um but high value a lot of activity a lot going on um and it has it has this like this

proof of work brain it's relatively simple and it works um so the to do this in a sane and conservative fashion uh the consensus the new consensus mechanism which will ultimately drive ethereum in the future is to be bootstrapped in parallel so essentially uh we're not going to disrupt this thing that we call eth1 the current ethereum chain instead we're going to allow people to stake and move real economic value into this parallel system which is that proof-of-stake mechanism which is called the beacon chain which is to be the brains of ethereum in the future and to bootstrap it in production unaffecting the initial system uh to

kind of get everything right to get stability to get the software and in production uh to test the economics uh in in real-world conditions etc etc um and so that's what we're doing at the end of the year is a boot shopping phase for the new consensus mechanism um if you're doing like the operating system analysis we're like you know putting the operating system in in beta and letting people test it and like they're getting early rewards for i don't know that's not that's not a good idea but um so bootstrapping phase that's like we're doing this in production we're putting proof of stake in production to bootstrap it to then take control of ethereum in the future um we throughout this year um and it's to a certain extent last year uh on single clients uh have been

iterating on test nuts um i don't know how many there have been there's probably been like three or four public multi-client test nets i might say multi-client because there are many implementations of this protocol which y'all are probably all aware of um so we have madasha madaya medalla whatever you want to call it it is a test net it has about a hundred thousand validators uh each validator is a 32 instance um and there's more than a thousand nodes there's kind of a distinction between validators and nodes we can get that later if you're interested um but about a thousand nodes on this network it is community driven um and community controlled meaning like the ef and client teams control

probably less than five percent of the stake on this thing um and so it's it's broadly run by the community it is not an incentivized test net meaning that um the incentive to be on it is there's no additional monetary incentive the incentive to be on it is to uh test out your software and kind of like show that this thing works uh and get your system hardened and security and and just kind of understand things before you move into mainnet um there's been some awesome things that have madasha uh you know some initial stability some craziness in terms of uh instabilities and and some uh bugs and hardening that happen and then um a long stretch of stability recently we've seen um

about 50 of steak um leave i think a lot of people are at the point where like i'm ready for net uh i don't really you know if i'm running on a cloud box like i don't really feel like paying for it if i have it at home like you know screw it i unplugged it from the wall i don't i don't know all the motives there but um plenty of people have turned their nodes off so madasha's actually in this um this time of no finale you need two thirds on the network uh to finalize and so it's in this like somewhat stressful state where the chains being still being built but things aren't being uh finalized and so you can't like aggressively as prune and things like that so um what we're experienced with madasha is like actually what i would love to

see which is like one of the more stressful environments you know 100 000 validator test net um in a pretty stressful environment and we're seeing some awesome prs and releases uh if you follow the client development um you know on better pruning strategies in this this time finality it's highlighted you know a couple of um i think uh i i noticed on on lighthouse uh proto and i were looking at something and noticed that like the attestation pool was like you know had four hundred thousand anticipations in it which doesn't make any sense but uh you know which is uh it could lead to memory spikes and stuff so um point being is madosha still exists uh madasha's in this like very stressful state and um gives us a chance to harden clients and and things before we go to mainnet

additionally we've had these like test net dress rehearsals recently which is one of the one of the more difficult parts of this process is really like that initial bootstrapping it's the making the genesis deposits and kind of coordinating all the clients at the same time uh to kick off genesis so we've had a couple of these like test net dress rehearsals the last one went really well um and so we're kind of in the final preparations for mainnet uh because of these positive signals from test nets and at the same time uh tying up various loose ends um which we can get into that's the status that i don't know it says that does that cover the testaments no that's super good that's fantastic and one of the things you mentioned is

how these test sets are not incentivized when we call like bitcoin and ethereum we call these things like crypto economic systems or at least that's what vitalik calls them that's really the only name that i can really think of that really gets on to what these things are crypto economic systems but right now what these test nets are are just crypto systems because they don't have the economics built into it so with to the best of your ability danny can you kind of tell us how how have things gone differently because of the lack of incentivization and maybe how do you expect things to change when phase zero rolls out which is like quote-unquote an incentivized test net that is fully cryptoeconomic there's a few things going on here so one like

first and foremost you bootstrap it off of gourley instead of mainnet eth uh if you ask a gorilla well they'll give you you know ten thousand each no problem um and so the ability to kind of come in and largely affect these systems to like be a large player in the in the test net is pretty easy um and because of that uh people that don't necessarily like i don't have much care other than like my personal want to understand and participate in the system uh to like keep the value that i participated in uh running smoothly so that's that's definitely one of the differences another difference is that we've seen and why we did a couple of these test night dress rehearsals is

that um like in madasha we had initial instability during the first like 40 epochs or something in epochs like six and a half minutes um and primarily the reason was that some people that chose to be whales in this test net uh didn't turn their nodes on right uh i won't say i won't name names uh but you know somebody that was like 10 stake was like oh [ __ ] i thought genesis was an hour later and like that can happen in mainnet but like i think you're probably gonna check double check triple check uh costs also like you can turn your note on two days before genesis i don't know why your nerve wasn't on you know so anyway um so that kind of stuff happens and that's

why we did a couple of the dress rehearsals to really like get uh get a good genesis going um and then otherwise like if i'm running on a main net node um i have real money at stake and not only can i make money but i can lose money um and so if i don't have my validator i don't have my node running optimally connect to the internet um you know if a bug arises in my client i'm gonna like dig in and like figure out is there a new release should i switch clients like i'm gonna be making i'm gonna be pressured due to the uh the loss of capital um and my lack of ability to gain capital uh to to fix things and so what we've seen is um i mean sometimes

there are client errors uh there are fewer and fewer but uh you know if 10 percent network goes off because of client error people are like slow to fix it um and when people don't want to participate anymore a lot of people are just like turning their machines off which if i were on main net i wouldn't do i would exit right like if i'm like sick of this like this is not for me i'm gonna exit i'm gonna take myself out of uh participation and so we see a lot of not quite uh rational activity if this stuff actually had value um and that's one of the reasons to go to mainnet and to bootstrap this thing with real production value is to like see how the economics work and personally so do you think when this is deployed to

mainnet danny some some people have called this like uh an incentivized test net almost because you know there still could be some issues post mainnet do you agree with that lens so i understand that framing that's not really my framing um i think the major difference is between an incentivized test net and what we're doing is that um the intent of going to mainnet here is that it is continuously although it will be upgraded and change in form uh that chain does become mainnet that chain does become the history whereas in incentivized test net i would say um has no intention to ultimately morph into what maine that is um and so it is it is

early um there's a little bit less there's real capital at stake there's less at stake is then there's probably less surface area to attack at the beginning and maybe less incentive for an attacker to come in um and so it does allow us to have training wheels on to a certain extent um transfers aren't enabled at the beginning which like in the event of catastrophe would allow us would allow the community uh to make maybe hard decisions on on reverts and things like that uh but so they're you know training it's like training wheels main net rather than um incentivization in my opinion i know that others call it incentivization i think there's validity to that claim but i think there's a definitely difference between a main net you know an early main net that's going

to evolve into the full thing versus like a test net that like is incentivized and ultimately won't turn into the main net yeah i got it all right so we so we went through all the test nets i think the big question on everyone's mind is what's next what's left to do here correct um i mean plain simple like the main thing to do is to release the version one with specs the community to choose a deposit contract mainnet canonical address and a genesis date there are a couple of reasons that uh myself and client teams have not done that final release of specs that would include what we would uh

call a canonical address and date um and the primary being and i haven't talked about this yet um i've been meaning to put it down in paper in a blog post um i've also just kind of been monitoring the situation but the primary being there's this crypto library it's called blst blast it's built by supernational it was um created actually this year and it is a highly optimized um bls crypto library and it's awesome you know like 2x gains on the fastest thing and it's actually easier to read and reason about and there's like ongoing formal verification of it this library is critical to

creating keys signing messages and critical in early phases and that if you use this library to generate your keys they need to be secure if they use it to generate your wallets they need to have like tightened uh good randomness and stuff um and if you're actually signing your deposits which has a signature associated with it it needs to be correct um and so this library is currently under audit um it is also under uh ongoing formal verification which has some initial results that steps forward and i can be ready until q1 uh but it is under uh security audit within cc um it's about two weeks then and we have another about two weeks of this audit to go um given that how critical this library is and

given that if there is a fundamental error in this library we could really [ __ ] some [ __ ] up uh in terms of genesis deposits um that is the blocker there are a number of other things going on uh clearly clients or clients are hardening uh clients are tying up loose ends um there's been a lot of like active engagement on medasha with some of the things that users are seeing in terms of resource consumption and stuff so that's ongoing there are some ongoing client audits and things and there are some loose ends being tied up around like user experience on the launch pad and other various tooling around deposits um but those things are all going to happen and are happening kind of in parallel uh but the the

the big blocker um on kind of finalizing the last little bits is this audit so danny there's a few questions with regards to the canonical deposit address uh how how do you foresee people figuring out uh which one is the canonical address and why why don't we just have a clear and cut answer as this is the address um i mean for one uh who can dictate such a thing um this is likely going to be signaled from many different angles from uh from the community kind of all at once uh i think that client teams ef um large community entities that have stake stakeholders in

the community um are planning on kind of coalescing on a single deposit contract um and signaling that all at once um to um one kind of solidify like the community has chosen this but two to also publish this information simultaneously kind of all over the web and it can't simultaneously campaign that says trip check double check triple check this before you send any uh any any money to it because at the end of the day the bootstrapping of uh the serenity the e2 consensus beacon chain um it happens with no changes to each one and so this boot trapping mechanism like the gateway

to enter into this consensus mechanism is a contract on each one and so i can deploy the bytecode you can deploy the bytecode anyone can deploy the bytecode we can verify that all of these are like the bytecode that we all intended to use but if we don't all choose the same one you know there could be issues with loss of funds and other things and i think maybe even more concerningly i can take that by code and modify it and maybe put some sort of secret withdrawals in there obviously if you're verifying the bytecode um you couldn't create like a phishing contract but you could create like a contract people are just wasting money uh but i'm definitely like we're all kind of concerned about fishing and and so we we do want to make sure that there's like a a good

community push to kind of like coalesce very clearly on this um you know i think this information will end up in that eip that i wrote up wrote up it'll end up in the vu10 specs on the spec repo it'll end up in many blog posts hopefully like block explorers and different things will also all be kind of publishing and giving names and and access to this information i think that's that's pretty interesting that for just a brief moment uh ethereum really is emphasized by the social layer rather than the code layer uh when it comes to like uh determining which is the quote-unquote real deposit contract i think that's a pretty pretty interesting feature and we you know any time there's a hard

fork um it's not there's maybe not a lot of risk in those but like that's a point in which the social layer of ethereum goes yes we're doing this right and like if it's adding like a bls pre-compile and it's adding uh you know it's fixing gas pricing that's not like the social layer doesn't have to like signal super strong but this because it's it's it's not a hard fork right it's a deployment of a code and and pointing to a code in this layer so there's a little bit more opportunity for um confusion and other strange issues that might arise there so the the bls uh blocker may be news to some folks who are listening to this and and watching to this uh watching this

and but let's take kind of the happy path so audit's done in like two weeks uh bison trails who's a staking company in the space they're you know like doing stuff with these two they wrote a post that said we expect the deposit smart contract and more exact dates to be released sometime in the next two weeks they they put this post out this week we still expect from mainnet to launch in mid to late november um i guess you know the question to you danny is when deposit contract let's take the happy path where bls uh goes well the audits come back everything's clean what dates are we looking at happy path is a couple weeks and

a minimum genesis date in 2020. very cool wow all right and we've got we've got we've got some holidays to navigate too is that on fixed minds um to a certain extent like yes but like so we have we have some uh some ability to select with uh this parameter called min genesis date right um and that is like a unix timestamp that is the earliest genesis can happen if a minimum number of deposits are hit um and so there's a little bit like if we assume optimistically that like deposits come in and we hit we hit it no problem

um then like we do kind of get to select a date and that would probably try to avoid any uh major holidays but if we don't hit those numbers then um genesis date genesis time is like totally up up in the wind um and uh you know hopefully hopefully that's not what i have to tell my wife that we're doing on christmas morning or something like that yeah whatever you know i i i take that back i think she would accept it as like a pretty reasonable gift yeah absolutely well okay so um you know the interesting thing about that i think is like this is going to be probably a light holiday year for a lot of people too anyway right so um at

least with respect to travel so maybe not all bad news if that happens here's a question from somebody in the community is asking to you to something you just sort of addressed danny how fast you guys think the deposit contract we'll hit 500 000 eth and maybe just give some background onto why that uh why that 500k eath is an important number right uh so do the background first um so essentially that is the minimum amount that can uh kick off the chain if more deposits happen before that minimum genesis time uh it can be kicked off with a higher amount um initially this number was put at two million um

find the balance between being able to bootstrap the chain but not have like a whale control the entire thing um obviously there are people with five hundred thousand eth there aren't incredible amount of people with 500 000 but there are players that could potentially game that and so the other the other balance between having an enough eath to kick it off and be in kind of in a secure zone is to in these early phase of e2 um there are there are risks that wouldn't necessarily exist in say two years time there is kind of this unknown uh lockup which as eth1 gets integrated into eth2

um you are then able to liquidate uh you're validating eth uh but until then you're kind of like you're granted to win it you're you're you're a validator and you can exit but then your ethers is sitting kind of in limbo and we don't have a date on when that happens so it's kind of an unknown variable right right right so we do expect like lower much lower participation early on and much higher reward than say in two years where um i i don't know say in some amount of time where the system is hardened uh the system is hardened so there's less technical risk and uh there's easier paths for liquidity so we do expect like there to be some amount of barrier to

everyone participating with all their eth right um and so given that um the low 500 000 um is kind of like in the event that it is not incredibly enticing to be one of the bootstrap validators um it's definitely an achievable number um and so it's it's it would be in a much lower security environment um obviously ethan's worth much more than it was uh maybe when some of these decisions were made and heath is worth like 90 um and so 500 000 isn't like a crazy low security environment but um it's really kind of finding the balance between what is the minimum that we can really expect in this uh high risk phase um

balanced with what's the minimum we can get away with in terms of security to to kick this thing off like obviously 20 000 ethan's is not not a sufficient number yeah but it's a little bit of guesswork um because of a lot of the unknowns and in terms of the incentives and who's gonna show up on day one i guess you know throw some numbers out here so i don't have the calculator in front of me but if you've got 500 if there's 500 000 eth in there like rewards are gonna be a lot higher so we're talking definitely double digits right maybe upwards close to like twenty percent ish rate yeah i just put this in that eip 2982 so let's check it out yeah it's like twenty three and a half percent okay and a half percent denominated which is

important not dollar denominated yeah sure yeah yeah i mean uh who like i finished that sentence danny i know where you're yeah do it do it who knows like in terms of dollar amount like what is that even like oh yeah yeah that changes every [ __ ] but like okay but here's so so here's what's interesting so 23 on the if it's close to 500k right so and that would be sort of threshold to booting this up we have uh so you know uh 500k ether right now price is what 410 ish eve price so that'd be like 200 million dollars parked in the staking contract basically

we have 8.9 million each locked in d5 that's like close to um how much is that it's like eight eight percent it's eight percent and that is billions of dollars um yeah it's i mean we've got billions of dollars locked in d5 and the ethan defy is making like i don't know two three percent yeah um not even right i thought you could make like ten thousand percent well that that was the summer and that yeah that didn't last long and that's on your stable coin it's not on your ether oh no yeah you don't get eat the nominated returns are the thing so um i don't know if we look at defy history uh at least

for me it feels pretty confident that are we gonna have people stake eth in eth2 for 23 youth denominated gains there's definitely going to be a subsection i feel like we're going to easily cross that threshold but it is because the the risks obvious the risk like smart contract risk is high right especially when you're participating in a new system and there are risks uh in eth2 right like so that's not just a free you know you don't just get a free 23 uh you have to run your node there are like technical attack risks they're all all sorts of stuff um and so the and and early on there's just more unknowns about like how stable this thing is which clients

are the best to use and that kind of stuff so like that number accounts for uh additional risks um but in the long term you know when say eight ten million each is in this thing um as we expect in the long term uh there's there's gonna be a lot fewer risks uh because this system will have existed in production and because a lot of these like the clients i think will be much more hardened by then and liquidity uh much more easier access to liquidity exists uh you expect a much lower return so it might look eventually like those meager d5 returns but uh but then it'll be much more of a

risk-free type of or low risk lower to those d5 um yeah it's a really nice property that i'm not sure if folks are aware of because uh feels very scalable on that so danny as this deposit contract gets up and running i'm i'm seeing kind of uh different parties that uh are relevant here there seems to be like you know people that are are going to take on day one no matter what and perhaps these people are are long time like ethereum people that probably like had you know like perhaps even not participated in the genesis sale i would imagine vitalik would be part of this group and then there's the the group of people that i think are just going to throw a few eth in just a small percentage of their total uh value into the deposit contract

and then there's gonna be people that that do nothing uh is that kind of how you see this stratification and also if somebody just doesn't want really to do anything do they have to do anything at all ever right um so i mean that's a reasonable way to think of stratification i mean there's people that uh it just it depends on your stack obviously uh you know i don't expect most people to go all in on their stack regardless of the quantity you know if you have 128 eth you might run a validator if you have 128 000 eth you might run you know a thousand validators uh but i i don't it's someone's going all in obviously

someone's like been waiting for this for for five years is like gonna drop their entire stack in it and like it is it has no intention of selling any of their eats for the next 10 years and it's like i'm all in but i think you're gonna see much more of the people in that middle ground where like you begin to experiment uh you throw in a couple of validators um and get your toes wet and kind of see how the system evolves see how rates evolve see how participation evolves and then potentially go in more and that's that's kind of on the in on the individual basis but there's also there's a lot of other like complicating factors to to uh this engagement of you you mentioned bison trails there's a number of these like staking providers and institutional

players um which is good i mean there's there's a demand for this stuff so i'd rather there be a ton of them and it seems like there's going to be a ton of them rather than just like one or two but some people might choose rather than to participate at home might participate in one of these like larger pools um or you know if you don't have 32 you might choose to participate in one of these pools with um a smaller denomination um and then yeah so i mean that's that's the range of that and then the well your second question was uh what if people people they don't really care to stake then just want to get on yeah and that goes back to like one of the

early questions we talked about in like this uh what eth1 is is the existing ethereum chain the state and the transactions and kind of like the continuation of that um and the plan is that uh this this chain is governed that state is governed by proof-of-work um and at some point in the future will be a hot swap of the consensus from proof of work to this beacon chain uh proof of stake and things would just be uninterrupted your uh your applications your contracts and your eth will just exist in this new environment um and so yes if you don't do anything don't do anything um honestly like phase zero honestly this like bootstrapping of consensus is for stakers it's for people that want to get involved um it's you

know obviously like there are stakers there are adapt developers they're community members and like the people kind of spam the gamut but like if you're not interested in being a staker then like just let it let it bootstrap let it go um and enjoy the fruits of it um in due time this is gonna be pretty historic though though i mean this is the first time a light kind of a live network is attempting to switch over from proof of work to proof of stake i've called this like economically if like if we think ethan's money is a monetary asset monetary type asset this is almost like a initial bond offering for the ethereum you know nation because you're able to put that monetary asset in a bond stake it and receive some sort of reward and all of this is happening before our eyes

live it's going to be i think one of the most fascinating economic experiments um you know maybe in history um yeah so here's a reader question though so i think you know back in 2017 we we faced kind of a couple of years of what felt like uh serenity delays youth two delays the reader asks or the watcher is asking what's the most likely reason why eth2 effort continues to to feel like it gets stalled or delayed like why has that been happening and like how do we know it's not going to happen again either at this phase zero or maybe even like phase one because like we get phase zero but still phase one and phase 1.5 can until

we can start doing things with it why the delays right um so this problem and the set of problems um has moved from the abstract to the concrete over the course of many years this research effort the abstract has been ongoing since before the launch of ethereum mainnet and continued for many years um it had many false starts it had meandering paths and it took a long time to bring the fundamental research into a concrete reality that even at the end of 2017 was still happening um some of the fun the

fundamental research driving the core of the proof of stake in the core of the charting was still really getting out the door um at that point we were also beginning development um and there was optim optimism uh that um all the the hard problems were solved um but it wasn't really until this mid 2018 shift where the abstract firmly began moving into the concrete um but moving into the concrete first had to move into specifications um and specifications took us a long time because these things are complicated um and so there was optimism that the early specifications were kind of where we needed to be and that implementation could uh quickly

follow the specification time took more into firmly into the 20 2019 range and thus really the implementations um were getting started and expertise engineering expertise in this domain was being built but we weren't at the point in which this stuff could be shipped really once the specifications solidified um and implementations could really bite in and dig in um in mid-2019 uh did the stuff concretely begin to move out the door um and what we've seen is client engineering teams first and foremost

understand this stuff and figure out how to build and manage it um it's one thing to write it down theoretically it's another to actually create sophisticated systems to do so what we do have now are sophisticated systems that can handle proof of stake and handle sharding and these that's what's being kind of finally battle tested and moving into production unit this year so there's a couple of things there there's the the abstract of like actually just abstractly can we solve these research problems there's writing specs down and then there's like the actual building of uh production systems um i would claim that on the abstract side on the research side um there are not a lot of unknowns um we do have very

firm grasp on uh the reality of the system that we're trying to build um and on the specification side we do have clear paths to production grade specifications on phase 1 phase 1.5 this stuff is still a bit influx but is in a place where we're kind of hardening and refining rather than coming up with the basics um we've also gotten better at writing specifications uh we have like a framework with how we write specifications we have like standard ways to write these things and standard ways to test and and and validate what we're writing um so those two the research stuff is not a big blocker uh specifications is ongoing um but we're in a pretty good place and on the on the actual engineering side um

engineering phase one in engineering phase 1.5 it's hard like plain and simple this stuff is hard any any amount of work in any amount of iteration is hard but the foundation um the foundation of these systems and the foundation of these engineers and these teams um is laid and so when you're talking about expanding say the role of what a validator does which happens in phase one they have additional duties not just the beacon chain but to these shard chains um this is an extension of uh what is already robust software there's already like generally frameworks for how to deal with validators how to kind of shuffle them around how to have different duties and store what they're doing and manage caches and shuffling all sorts of stuff so a lot of the

fundamental problems are solved it's really extending the ability of these systems and so i will say this time and time again this stuff is hard and like fundamentally like that's that's the answer that's why uh this that's why there are delays uh this stuff is hard it's never been done before um but we uh the researchers the developers the testers that everyone kind of involved in the system are much better positioned to move um professionally and move quickly on subsequent phases and so yes there is complexity yes i will not give you a timeline uh but um the ability to extend these systems and the ability of the engineers and the

teams you know is up there awesome thanks for that danny one question that we have coming in from the you youtube is uh what would you like to see from the client teams uh you mentioned loose ends are being tied up one thing personally i'd like to see is the new client switching schema i actually don't know what that is maybe you could help elaborate on that right so one thing um i'm going to beat the drum on client diversity all day every day i think that we have a little bit unhealthy distribution the prison team has been around for uh for longer than any other team and they've had public test nets for much longer than the other team and so i think a lot of the community

is kind of anchored on this client because they're just familiar with it it's a good client uh but there are other very excellent clients um and so what we saw in madasha was we saw early madosha we saw a critical error um with how prism was handling time uh which caused like some madness uh it also eliminated us the fact that like prism was probably 70 plus percent of the network and during that we could have seen a much quicker resolution of this error if validators were able to cleanly simply and securely swap clients but there actually was not like a standard in place for me that will cleanly do that and actually if i

naively just swap my client due to the nature of the timing error on the prism client i might actually been uh liable to be slashed um if i didn't like port the information of my previous messages to this new client and so one of the great things there's a lot that came out of that midasha incident um all i think all clients are like 10 times better since that incident because they got to deal with like a crazy stressful forking scenario with like 10 percent of network that remained live um so that's good good in and of itself but another thing that came out of it was there's this um validator database like interchange format essentially if i'm running uh prism i can go you know export uh validatordb

um and then i have like this file and then if i want to run lighthouse or nimbus or teku i can just go import validator db import keys or like import import my keys and it'll say do you have an existing db or something and you give them the file i don't know what the ux around it is but this standard exists so that i can safely um and easily kind of port between between clients um and so that's that's what they were mentioning there um and this um there is a standard uh there is a standard that is implemented i know by lighthouse and i believe by teku and maybe by uh prism and the standard uh is being prepped to be you know an eip erc um and so

that's that's one of the kind of things like when i say loose ends that are being tied up um so that um these you know people not only can run their clients but in exceptional scenarios can can run and manage their clients another thing that i it's kind of funny it's like um i think teku like recently merged a pr that actually lets you submit exits right like that's that's like the tail end of the validation process and like not something we're like consider like that you um tat like when you're when you're imagining a test night you're usually like genesis and like run stably um but because of that you know they had this like feature just kind of like hanging in an issue um of actually being able to

submit an exit you know and so that's that's something that like when i say loose ends uh that's not a complex feature it's just sign a message and broadcast it uh but it didn't exist um and so that's that's the kind of stuff um in addition to the final hardening and kind of security preparations that are happening one thing i want to go back to the the switching the the client switching schema scheme i guess uh historically we've had instances in ethereum 1.0 where one client got ddosed right and so it was this was the shanghai attacks way back when and if and you know there was basically if i'm remembering uh correctly there's

basically just two clients at the time and one client went down and the reason why a theorem didn't go down was because the other side of ethereum which was the uh uh what was the other client that that there's parity and gathering parity they actually both independently got attacked but at different times at different times right and so like they said one was the backstop for the other and what you're talking about with these switching schema uh switching validators in ethereum 2.0 is that we're make we're trying to find a way to like just uh switch at the press of a button right and so for people that's trying to attack ethereum it's like playing whack-a-mole because what these clients are trying to do is they're trying to make such an easy way to switch on to a new uh client that may not be succumbing

to whatever attack you're trying to have so it's just about producing anti-fragility this is all correct right right right and and some of this we get just by being in a multi-client paradigm without active switching right so if um a client that is one-third of the network or say a client like 25 percent of the network and it gets attacked and everyone's takedown taken down the chain can still be built and it actually can still be finalized now take a client that's 50 of the network uh the chain can still be built if that client is taken off but it can't it can't finalize so it can still provide liveness and some quality of service but it can't provide provide that uh crypto economic finality um that we would desire um and so that's that's another component like so

if there's a 50 client taken off and there's not a good way to switch then we're kind of in this like limbo state where we're not finalizing and so this adds additional um ability to um rectify scenarios and personally you know if you're if you're a validator and you just want you want good uptime like if you have issues with a client you want to be able to switch regardless of whether there's an attack um so this just kind of provides that baseline um ability to to be able to manage and kind of control your own destiny and that that's that's i kind of alluded to the fact that i think that prism has too high of a community uh share right now uh because um we have we have four clients um four very viable clients that are that are

going to mainnet we have an additional client load star with a javascript client they're awesome they bring incredible javascript tooling but they're not quite ready for mainnet um and these four clients um if we had say like every client had 20 to 30 of the network um then we'd have like this kind of built-in redundancy where if a client goes down uh the liveness remains the the finality remains live um whereas if we have a client that has like a huge percentage of the network they just it embodies like a huge amount of uh much more risk for both the users uh the the validators and kind of their profits um as well as um this finality mechanism of uh of the protocol so again like being able to switch easier helps um in the

event that a majority client has issues um but i'd also love to see a better distribution in general very cool and we have we have four viable clients so like use them maybe maybe we'll see some of that uh particularly when we get to you know production maintenance and real money is at stake and people want some redundancy against one uh client outage so let's hope for that here's another viewer question so after phase zero how long until phase one and two and maybe let me just slip uh phase like 1.5 in there conceptually realizing this is all part of serenity uh of course but uh what about these other phases so give us happy path maybe yeah i mean i'm

um these client teams uh they have uh they're they're much larger than they were a year ago uh they've been encouraged to kind of like hire and build up the expertise but they generally have like some like very expert resources that the leads um that are driving a lot uh the goal and the intention is to get phaser out the door um and shift a lot of these expert resources to sprint on phase one um it's been very difficult to prioritize uh the development of phase one by these client teams while they have uh the impending uh launch of phase zero uh but phase zero will go in more into an iterative uh maintenance mode uh for these clients um and the expert

resources will spend day and night working on shipping phase one i think that it's realistic to see that next year but i need these engineers to dig in come you know january 3rd and begin to flush it out again a lot of the core of this system and a lot of the complexity of the system is the managing of the speaking chain of you know hundred thousand hundreds of thousands of validators and consensus entities and caching and um databases and all sorts of stuff and a lot of that is in place um and a lot of what phase one is kind of an extension of a lot of those basic building blocks so

i'm pretty confident that uh you know we'll get significant kind of basis in place early next year and test that's going uh but there's a lot of unknowns and so so looking at that so uh phase zero builds the brain right so we've got this whole body that is ethereum it's got its own separate brain now we're building another brain over here but it doesn't have a body yet right so it's in kind of this lab and then phase zero launches with you know and we're building the brain which is kind of the consensus engine right what additional benefits do we get with phase one we had vitalik on the podcast and he was talking about hey it's kind of like a um sharded data availability layer right for us so are there any winds there

and then uh phase 1.5 it's it's kind of my impression that that's the point at which we can we can delete the other brain and we can merge the body and like the the bodies together into one thing but what are the wins for phase one yeah so um so the goal one of the primary goals is to migrate the existing ethereum chain from proofwork for mistake um in the meantime while this proof-of-stake mechanism is being bootstrapped and built embedded in production and kind of iterated on we will also extend that proof-of-stake mechanism to come to consensus on a lot of data it's sharded data and provide a whole host of extended data layer one data availability

capabilities essentially um we expect this consensus mechanism to be able to come to consensus on anywhere from like one to four megabytes of sec of data per second and um that's what phase one is is taking that consensus um and uh and getting it to come to consensus on a lot of data and data's kind of this weird thing where like uh i'm aware of how important layer one data can be to uh applications and things but i think maybe the broader community is just beginning to kind of bite into what layer one data even means and what what what it can do for us um and i i distinguish data from data and computation so like layer one

data in phase one doesn't really have like a layer one meaning it's just the ability for the consensus system to like make strong cryptoeconomic claims that this data exists and is available so what does data do for us if we don't have computation well i think first first and foremost what we expect to be used for is uh is rollups uh which is like one of the main scalability avenues that ethereum wants to take in the short term but also probably the medium and long term um where the amount of layer one data dictates how much these systems can scale i think you can get a couple of order magnitude with just the data of the single ethereum blockchain today

but with you know 100x that in a sharded solution we can take those two or two doors of magnitude and multiplying by another couple um and get some pretty incredible scaling um but another interesting thing is like what is what is layer one data what what can we do with it you know if we don't put computation on it i don't know like there might be all sorts of crazy stuff like just having this extended capability one i think is going to provide us like quite a bit of scaling through rollups but um there's all probably all sorts of decentralized applications and things that we haven't even thought of that um all of the innovators out there and all the people that like always want to play with the new gadgets um are going to take an extend um and so layer one data we expect to just be an asset in and of

itself um and so phase one even though the existing uh proof-of-work chain what we call eth1 isn't integrated back into the system it can learn about this system through a light client on eth1 and have access to make proofs about this data availability layer and so all of these roll-up mechanisms that are in production today and kind of on the verge of being in production in the next quarter or so um can not only leverage the layer one data of the existing ethereum chain but this highly scalable data layer of um kind of the beacon chain system and so uh we do expect to be some pretty major gains i expect this to be a big

pretty big asset to the community at that phase um and then subsequently um we'll have the integration of the existing ethereum chain into this proof of state consensus um for a little bit more native access to that data layer all right so um we've gone through kind of all the phases right so again happy path maybe uh it looks like happy path says that the the uh contract will ship and staking will ship this quarter and then happy path would maybe say phase one ships next year i think that maybe the the last question the one to wrap on uh danny is when will the old brain die when will we finally be rid of proof of work for good

give us another happy path on that because that would make me very happy uh when it's ready um there is so there's been a lot of work done on this um mikhail from txrx at consensus and guillaume from the geth team at the ethereum foundation have worked on a a prototype of the basic uh the merge the beacon chain controlling the consensus of existing ethereum um it's great because it reuses uh most of the software sophisticated software beth1 geth etc and it reuses the same uh public interfaces and so you can take existing wallets and stuff and just interact with ethereum but in this new environment uh so we like used metamask to send transactions to um you know an ethereum chain controlled by the beacon chain

very cool um in addition to that um there's now there are probably three or four of us uh built working on this merge um and now there's upwards of like 10 people um there's a lot of different things to think through like shard data fee markets um actually how you handle security at the point of merge you know when uh the proof of work uh dies and the proof of stake takes control you know at this inflection point like what are the things to think through uh we're currently prototyping that inflection point um and a number of other things like at that point you have the ability for native validator withdrawals um and more native deposits um and so figuring out the mechanisms of

uh withdrawaling and withdrawing contracts and all that kind of stuff um that stuff is all in the works and that stuff is all very independent of this phase one development um and is happening in parallel and kind of picking up steam um so i think that by mid next year uh we will have like pretty robust prototypes and beginning to have pretty rough specifications of the different uh dimensions of this problem um and although i think client teams will be uh digging very deeply into phase one at that point um we'll be ready to really plan and begin for uh serious planning of the merge um so you know

take it take it take it for what it is i i like i have like we all have like an optimal path and like i probably have a better understanding of some of the problems at hand than others but these are deep and complex problems and there's a lot of different moving parts and i don't want to i don't want to put engineering teams in the position of having an arbitrary deadline that i happen to say on a podcast i'm not going to say it that is completely fair uh thanks thanks a lot danny yeah danny we want to give a big tip of the hat to spearheading this entire entire effort so you know thank you for for everything that you have done and and you know pushing the

boulder up the hill uh so far uh and so thank you i get the chance to talk about it and i get the chance to work with kind of in the middle of a lot of it but like tipping my hat to uh all the incredible researchers at a theory at the ethereum foundation and elsewhere um and like the like the engineering talent on this project is is unbelievable um and they're really like the the unsung heroes this process so shout out to them yeah should definitely shout out to them danny if there's one thing that the bankless nation could help with this effort what would it be what what could uh listeners and viewers do to to make sure that you know ethereum 2.0 comes alive in the way that we want it to

i mean phase zeros for stakers um if you are planning on getting involved uh get involved today um check out the e-staker community dig deep run some software run some non-majority client software help yourself and help the network uh by using a diversity of software um so i mean that that's that's the big shout out um if you're planning on staking and i mean otherwise if you're just in the ethereum community or you're building daps like keep doing what you're doing um we're trying to make a better future for you but like there's a lot of incredible [ __ ] going on today so like if that's your thing just stay heads down and keep making ethereum awesome absolutely yeah okay and if you guys are interested in doing that i would actually super recommend the eth online global event that's tomorrow there's

going to be a talk about uh staking for beginners uh at uh 17 edt or 1700 hours edt whenever that time zone is you can check it out at eathglobal.online there's a schedule there a bunch of other stuff too as well such as eip1559 and if you're interested in a lot of the content that danny was talking about just now you're probably going to be interested with all that content going on there so learning more that would be a great resource danny thanks so much for for this ama really appreciate your time thank you it's my pleasure talk to you all soon cheers thanks a lot