New Ethereum talks, every Monday. The week's conference uploads by event, in your inbox.

Loading player…

Show and Tell

DevconSun, Oct 7, 2018, 12:00 AM

Visit the https://archive.devcon.org/ to gain access to the entire library of Devcon talks with the ease of filtering, playlists, personalized suggestions, decentralized access on IPFS and more. https://archive.devcon.org/archive/watch/3/show-and-tell Virgil Griffith gives their talk titled, "Show and Tell" Speaker(s): Virgil Griffith Skill level: Intermediate Track: Developer Infrastructure Follow us: https://twitter.com/efdevcon, https://twitter.com/ethereum Learn more about devcon: https://www.devcon.org/ Learn more about ethereum: https://ethereum.org/ Devcon is the Ethereum conference for developers, researchers, thinkers, and makers. Devcon 3 was held in Cancún, Mexico on Nov 1 - 4, 2017 Devcon is organized and presented by the Ethereum Foundation, with the support of our sponsors. To find out more, please visit https://ethereum.foundation/

Transcript

so hi everyone so originalist was called gas pricing and so we have about six months ago and I was to show you a thing on how we're gonna probably different gas assets and short gas has to be idea how to price them insure it like we basically we look at the screen of prices and we say oh that seems a little low that seems a little high we just kind of manually tweak it up and down so we're trying to make that better but we don't have it ready yet so if you were focusing on Casper so so so I was showing you sort of a list of cool things that we've been working on for the past six months and hopefully we'll have gas pricing for you next time so I'll kind of show you we got so far so really the big change in gas that we want to change this thing isn't called rent so there's a big problem this is working okay wait yeah so there's a big problem is that so right now in aetherium storage and in bitcoin as well when you you'll pay once but you get the storage forever and in short that's just not sustainable and so right now in bitcoin is it's not that big of a deal because the amount of storage they're doing is just so low the problem is that in theory and people can make extremely storage heavy contracts so it makes it makes a lot more possible people just to break this like small economic misalignments make it people can like really exploit them on ethereum so ensure we have to be a lot more aggressive about this but anyway so the idea behind rent is so when you when you put down a big thing of storage you may see pay per block and if you and if you eventually run out of money your contract gets deleted and yep I know that kind of sucks and I know all of you are gonna hate it but you know what are you gonna do so so so we do have one other idea that might actually solve it and Vitalik we talked about today I recommend it brand new idea there called stateless clients um do I have it oh well here so okay the cold code called the stateless clients it's sort of a sort of a brand new idea basically so base with this you would need me any rent at all but the but the size of transactions goes up about 20x and there are some new DDoS vectors we don't quite know if this is worth it basically the idea it basically depends on which which one increases more quickly so BAM so we did some calculations here bandwidth doubles every three and a half years as hard drives double a little bit faster but storage IO doesn't seem to increase much at all or at least linearly and so we look at these and we say oh well you know well well it'll be harder to daus this there will be two daus that and so therefore wouldn't so therefore we want the big fat stateless transactions anyway if you want to learn all about this it's these maximally like clients it's sort of a new idea were we're still talking about it but but in short you're going to either deal with these these these like clients or to deal with rent so uh or think about it if you have any any data on on how storage and storage IO increases in time that would really help because you know we want to know like how hard this thing is is is to das all right that's the first thing the second thing is Casper you when you hear about Casper all friggin day in short Casper's great so here's the great things about Casper I mean I I knew like proof of stake doesn't I mean people are often a little conflicted about it nah man proof of stake is friggin awesome so what the best thing about proof of stake is that that you gain a vastly bigger stick to beat people with so here's the idea so in Bitcoin miners miners they behave because they want to get future rewards from the from from the from the block wars or and but the more their advocate behavior the more they can do this and improve his stake it's much more explicit it like like like it gets like like like it gives you like um like it like you can imagine it's like a it's like they do something naughty you can like beat them with like it was like a million dollar stick like immediately and you're just like you know and it's like and you just can't do that in Bitcoin like like you know so you could imagine so like a block reward is kind of a carrot or you can like you know make the cow kind of come your way you're like oh isn't it great you know but with the proof of stake you kind of have a ring around its nose and you go yeah you know you can like real to do what you wanted to do so basically steaks great oh we also have vastly lower inflation rates so in proof-of-work mining the biggest cost are both your mining rigs and the are your electricity and and proof of stake both of those get set to zero and you're like yeah so now we can make pay you like like ninety nine percent less you know and they're gonna put up with it because because because because their costs are just so much lower so you know it's great so so so this means a lower inflation rate for all of us and we all love low inflation low inflation rates the other one is that is sort of it sort of it kills mining cartels while they're in the cradle so in Bitcoin let's have a trouble with mining cart with the basically the mining pools having different incentives than the dev team and like we all know about that one nice thing about this is is that in proof of stake there's no sunk cost into the rig so the problem with with them with with the Bitcoin miners that they've already put in the countries of thousands of dollars into these rigs and like they just can't I mean I mean they can't just change like I mean value no really they are locked in they are locked in so that's what they will fight tooth and nail one nice thing about about proof of stake is that there are no mining rigs so as a result you know you know if there's some some problem you're not locked in you can always go somewhere else you can always go invest in like a mutual fund or something anyway peace taiks great so yeah so so this is my most recent paper by Vitalik and myself in short we are like 95% happy with it so I know it is it is a hybrid proof of stake and proof of work model we want to switch to be foolproof mistake model we're not exactly sure how to do that yet but we have some ideas and hopefully next year we'll have one for you people often ask when is this gonna go into the main chain we don't know yet when it's ready basically but regardless it's getting increasingly ready and and we're and we're proving theorems about it I think there's there's a Casper test net coming up I have probably I don't want to play with it but in short this this thing is growing and and we think it's great like well we we see there's basically no downsides all right so social stuff so this is kind of fun to do doing along the side yeah fun fact ether seems to be more halal than Bitcoin so so so so Wow get well we want Caspar and things like that always reading about sort of the history of financial regulations very hot subject I know like riveting and an Islamic finance a lot of the rules are actually really interesting if you're just curious but anyway so one of the rules they have for their currency is that your currency has to be intrinsically valuable and yeah so they prefer things like like barrels of oil or like cellphone minutes over say like US Dollars and sometimes you even see Islamic finance companies you cellphone minutes instead of using say USD because because it's intrinsically valuable well you can like consume it and I say oh well ethers kind of like that you can used to power programs and that's different than Bitcoin and so I so I showed this to one of my I don't know much about Malaysian finance but I read a little bit about it I should to do to too much my Malaysian banking buddy and he's like actually be true I was just like wow really and so so yeah so so so we are currently in process you can get ether classified 100% Sharia compliant so you know if you're in your Sharia society yeah yeah so you can you can do that I'm here because this is just kind of musing so this is a new Bitcoin currency symbol so no one really likes the little bit a little little triple triple triple lines like the Chinese Sun no one likes that so I was gonna say oh well we make so by the way we have a new logo do y'all know about that yet oh we have a new logo this is our new logo basically same as the old one but it's it's flatter now but otherwise same thing but anyway so we're gonna propose this as a new new currency symbol for fruit for ether but unfortunate this is not yet in Unicode so we get a little black diamond so in short if you want to try a new logo you can use the black diamond all right so here's other things so we have a whole bunch of new research projects and we finally wrote down what all of our research does it's all right there and this is sort of a call for volunteers if you're in doing work in any of this stuff there's more details there you can send me an email about it and we'll see if we cook you up yeah we're open source saying people are can ask where can I contribute well that's where probably the cool stuff about this is is that you get to hang out that you to chat with myself and metallic and other another kind of smart people when people kind of like that so we can give you that all right so here are some fun things so I also work on the the etherium name service and I can so we always so we had the dot F thing so ensure what we want to do we want to make these accessible via the UM via the global dns hierarchy so you want to be to type in something like you know example dot e that have a type and be a type it into a browser and have it work the first problem um death is actually owned by Ethiopia so you can't actually use no I know I didn't know that either you know it is you know so we actually considered holding holding Devcon for in ethiopia you know so that we could get this but you know but they were secure I mean it seems to be a non-starter but you know hey if you all I mean I'm down for it if you all are so okay so so another thing is so we be pitches to I can we just have sent a representative to the big ICANN meeting it was actually yesterday it was in Abu Dhabi and macey basically have really strict rules about a based upon known take a take away mow taking away domains so if you get coca-cola East coca-cola wants me to take it away from you and in short like if you want to be resolvable via the big by the big DNS namespace you just you just you just gotta accept this sorry just the way the world is and so we are proposing that we could have on multiple zones so we could have like we'd have the regular da teeth where it really is like first-come first-serve and then we have you know like like like dot compliant or whatever where where we comply with all the regular stuff like com does and so in but you can tell any different zones so if you want to have like a regular name you know in like in like dot anarchy you can have that and if you wanna have a name inside a nice place yeah you get you get visible one nice thing about this is so so so so we had a lovely meeting with ICANN a bit to my surprise they seem actually actually actually pretty chill with all of us doing this so they proposed that we get a place in the single dot ARPA you pythons seen DARPA before because there's only like six domains on it in the entire world but in short this is this is this is the the the Internet infrastructure testing ground so the ideas that you're gonna register you know example dot if DARPA and you'll be the like put stuff on it and have it resolved and go places odds are likely that y'all will probably do something really terrible there and then woman should be kicked out but you know but like please try and restrain yourselves because I mean just just restrain yourselves but really the good part is so for example so basically if we succeed there then we'll then we'll have a case that if we want to get something you know like like like like dot etherium or something like that which actually works in the big space it's much harder for me to tell us no because before they say oh we can't give this too big so you gonna break the internet and they say oh we're not gonna break the internet because we excessively graduated from your sort of DARPA testbed yeah so anyway so those those are kinds of the cool things there will be a few more announcements about ENS from Nick tomorrow at 5:00 recommend that in short you'll you'll be a register real domains on aetherium sooner than you'd think so but I can't tell you about it alright so something else I made so it so everyone loves Oracle's so I made one using using the using of the Wolfram language and I want to show that to you and you can ask a lot of cool things about it so let's just let's just do one so I was engine doing on prediction mark so you all use like Wolfram Alpha you've probably seen that this is like Wolfram Alpha on steroids so I'll show it to you can we do um can we do the Kimmy can you give me access now oh there I am hello that's me okay so check this out I guess I'll make this work all right here so it's a water school is floors click this thing here so you can do things like like what is the price of gold is the price of gold Dan September 2016 so Mike that you buy some work and it will go so okay so so you can input in free text my pet Wolfram language so here's the real magic here the real magic here is that is is is is that so then what this does this converts a natural language expression to a Wolfram language expression this mortar like this now becomes an object you can directly query so what you can do is you can take this little entity thing you can copy it and you can like paste it into here OOP and you hit the submit button what oh come on wait that should work right oh come on all right well let you out on this one let's just do when is the next next next Halley's Comet it should be to do that okay wait about working staff that should work all right let's do this one okay can we factor integer you shoes so we deflect a nerd all right well this is this as this is not going well all right let's try it let's try it let's let's let me try one more all right well you know serves me right for doing for for doing it I did do my 11th stage basically this will basically I'm off stage I'm gonna make this work in short and it'll probably work within like five minutes yeah I don't know I don't know I don't know what the issue is well anyway here's basically the deal the dealio here let's let's let's go back to the PowerPoint you mean can you go back maybe alright let me just keep trying this he's still working yeah alright well I don't know what the issue is okay okay so he's Jose alright alright so so so basically the gist is that um that's so so you can execute so you get these queries so you can put this factor integer and it will factor it for you like right there so you can like outsource hard computation so you can do something like you can ask you know you know what was the price so you can say oh like invert this matrix or interpolate this polynomial and it will just do it for you so you'd have you have a gas intensive computation you can you you you can just ask or you can just ask wolfram and it will tell it to you and we can also ask we also do on prediction market thing so you can ask you know so okay the cool is it can we still not get get get the PDF all right great all right so here's kinda the idea what you can do here all right so here's the idea so you can type in sort of a natural language query and you get this Wolfram entity object out of it and so and this and this is like a piece of code that you can type in and it tells you very precisely what you mean and so so and you can like look this up for anything so the idea is if you'd say oh you can change this things like dozen seventeen and put it like on the 12th and you say and you can put this and put this into a prediction market for the price of gold and that's that's like basically the idea so in it but it has has an extremely sophisticated well as it has a very elaborate grammar so you can say oh you know how many murders between United States and it has look oh guess what there's a query for exactly that and it has a query for like all kinds of things right and you can also ask you things like who's Prime Minister Singapore and stuff like that so I think so this is this is basically this gives you a way to number one query query outside data in a very generic way a number two key idea of outsourcing very high gas computation like inverting matrices or knowing cone umber is prime or factoring a prime number or interpolating a polynomial so you can give it data points on a polynomial and you can say oh oh what is a function that fits these and it will just tell it to do all right so so these are so these are the so these are the things that we're interested in now if you're interested in helping you out these are things that are great so we're looking for research volunteers we're looking for trends on storage and storage i/o and you can also use that you use the cool black diamond for your ether symbol and if you want to play with with Wolfram language you can do it at a war occult vertol dot gr and that should be working probably within three minutes all right I got [Music]

Automatic transcript — names and jargon may be misspelled.