Decentralized Asset Custody

Transcript

this is really exciting this is the first time i've spoken in like two years so bear with me i used to talk at you may know me from new cipher i was a founding engineer there i'm no longer at new cipher i work on entropy now entropy is my own project so this is our first time talking about it publicly we're really excited to share it with you so let's get on it so today we're talking about decentralized asset custodians and what let's talk about asset custodians first you may have heard of people like fire blocks or coin based custody or even gemini custody and an asset custodian is generally a trusted individual or organization that holds assets for

another individual or organization um they store cryptocurrency assets in cold storage mechanisms and typically leverage things like threshold signature algorithms like threshold ecdsa and like multi-party computation to avoid putting funds at risk this is the basically the core principles of how these mechanisms work they're centralized of course as well and they also tend to offer products that engage with defy and the wider blockchain ecosystems fireblocks does this they have things like some d5 primitives that i know work out okay and you know sometimes they offer staking

and things like that their users are generally institutions as well and other really ultra ultra high net worth individuals you know these are people custodializing you know tens of millions if not hundreds of millions or even billions of dollars of assets at one time um they they get to charge extremely high fees because of this because an institution generally doesn't care you know if they're paying 200 000 you know maybe a quarter to custodialize that much money so in like in short they're not exactly a crypto native entity this is a very regulatory locked-in industry so i get to introduce

decentralized asset custodians this is kind of a novel thing that we're building and working on and as far as i know we're one of the first and the only ones in the industry that are actually doing this i think this is actually my favorite slide um tag yourself i am the kneeling soldier that's me right there decentralized asset custodians what is this this is a trustless protocol for securely storing crypto assets um like i like to describe this as what amms did to centralized exchanges decentralized asset custodians or dacs it's not as fancy as amm i don't like the acronym as much so decentralized asset custodians

will do to centralized custodians um these are crit it's a basically a crypto native analog for cold storage mechanisms that we discussed earlier um you know they use threshold signature algorithms and other advanced multi-party computation and cryptographic um algorithms for that usually around consensus um i say usually because as far as i know uh what i'm about to introduce is one of the first of its kind um so it's kind of a i would say it's a big deal and so i'm kind of just giving an abstract and general overview of what this is and hoping maybe we learn more as we go on the difference between this and something like coinbase custody is that we're transparent we're decentralized

and community owned and operated um it doesn't it's radically inclusive for everyone regardless of net worth um you know you don't need to be a decent millionaire or a sent to millionaire to use a custody solution for your crypto assets you should be able to use it if you have 10 bucks of crypto or if you have 100 million dollars worth of crypto or a billion dollars thankfully we have cryptography so we can actually do this um so yeah well with further ado entropy uh please clap just kidding um yeah so entropy as i was saying is the first ever

trustless and i mean this trustless decentralized custodian protocol um you do not trust it whatsoever uh it's sort of like two-factor authentication for your private keys except decentralized um i like to say honestly it really does feel like a new paradigm for blockchain interaction we're pretty early right now but give me maybe a year maybe two i bet some of you are in this room are probably going to be talking about how you like to use entropy for interacting with defy we're built on top of two-party threshold signature algorithms i'll get more into that in a bit but to in short what i want to say is there is no risk to user funds if the

network gets hacked this is because of two-party threshold signing and more on that and it works on pretty much any blockchain this is because all blockchains generally use the same cryptography um and we're just able to kind of build a one-size-fits-all solutions for most blockchains um okay moving on so what does it do basically anything that you can relate to a centralized custodian for example what you use um fire blocks for or a coinbase cluster you may have like uh authorized signers you may have like you and your organization share a key or somewhere you know and you're able to say oh this so-and-so gets to spend this much and so and so gets to interact with these

contracts well we do the same thing we get to do address white listing so if you want to only interact with curve or if you only want to interact with unit swap or susie schwapp you can say that if you want to say oh my private key should only be able to spend ten thousand dollars in a week you can do that our network will not will not assist you in signing transactions that incur a spending limit that incur an extra amount over ten thousand dollars if that's what you set up another thing that's really interesting is sort of this um concept of transaction time locks which is something that we're really kind of

interested in doing uh on day one is you know everyone's worried what if my private key gets stolen well in our case if somebody were to steal like your share of this network and how it works uh if somebody were to steal your private key and they try to spend with it it's going to basically say hey you set a time limit maybe it's trying to spend more than a hundred thousand dollars or maybe it's trying to spend your entire wallet our network can see that and go hey let's make you wait a week before you can do that and we might need a second authorizer and that brings us to multi-factor authentication you know so maybe we even include your phone in on this and suddenly now you have like 2fa now if somebody steals your private key

share it's not going to work they're not going to take your funds now we get all this and we realize okay well we need it to be a little bit more powerful uh we need to be able to say it's really hard to continuously write rules and things like this so we realized it needs to be arbitrary and that's where we come in with this wasm thing we want to be able to implement arbitrary programmable rules via wasm so the idea is if you're a power user say like a fund you know and you want to be able to say hey we only want to interact with these contracts and have this time limits and who knows what else maybe some uh we'll we're going to try to get some oracles on our network as well so we can kind of get users access to that

so maybe you want to use that in your programming set so now you can say you can write your own code and your own rule set for multi-sigs kind of but except decentralized and with custodial so write it in wasm deploy it and now you have a completely sort of autonomous trustless custodial service and like i said all this is decentralized and i'm not kidding this is truly decentralized it it works this is not a pipe dream um so we'll get into it the cryptography we basically use two-party ecdsa and you might have heard this from lindell 17 and for any cryptographers in the room lindell 17 is the protocol that

we're thinking of going with on day one it's still we're still building some stuff and prototyping with respect to cryptography um but we're going to figure it out and try to see what is easiest for us to move forward as well so other people are using gg20 so let's be let me be really clear two-part ec dsa is not your standard m of n ecdsa basically m and ecdsa is when you have in shares of a private key and you have m signers on that share who want who agree to sign a message so for example you could say five of seven so you could say well i have seven shares and i need five of those shares to sign the same

message to produce a signature on a transaction the problem with using this in a decentralized network is that you have to have a staked amount of assets into the network and you can never custodialize more than that staked amount because otherwise there becomes a collusion risk so there becomes a huge problem there for us that's just simply just something we can't do that would not let us allow us to be trustless you would have to trust it to some degree because it can be colluded against so we settle on two-party cdsa what that means is we simply took one of those two parties and i'll point out here party one and party two we simply took one of those parties and we said hey let's turn this into a decentralized network that's what we did

so now the logic is really simple we say do both parties sign some transaction hash if they don't no signature is created and if they do a valid signature is created it's that simple um both parties have to agree on signing this message there's no way for you to sign a message that you don't agree with if i want to spend or trade ten thousand dollars worth of new cipher token to uh eth or vice versa uh if the entropy network does not agree does not want to sign that then the transaction will be made if you don't want to sign the thing that the entry network wants you to sign if for example there's a malicious actor or something you won't sign it it just won't work

this is very similar to the two-man rule you know with nuclear weapons where if you try to turn the key it requires both people to turn the key like i said again to reiterate functionally this means that if the network is ever hacked the worst case is denial of service never fund loss so this is quite novel for us i think and we're very excited to bring that up um so that's pretty much it for me thank you i'm tux we're hiring um we're looking for a lot of product people right now as well as some engineers we have actually a lot of time seven minutes left i kind of ran through this so i'm happy to answer questions actually i know it was a lot is there any question

no so maybe we'll you know awesome because we are running after schedule so maybe we can sorry i can't hear you we are a little bit behind schedule so thank you for yeah of course thank you very much