Hugh Karp: Nexus Mutual – The Decentralized Insurance for Ethereum #345

Transcript

this is episode our episode 345 of guest:you carp from Nexus mutual this is Brian crane and this is sonia agarwal and today we're talking with you carve of nexus mutual i guess you could say we have a little bit of a insurance theme going on right now because last week we just did an episode with open and today with nexus mutual we explore sort of a different form of how to do insurance maybe a model that's more familiar to

most people and what they think of insurance which is this mutual concept where people sort of buy in and pay premiums and then they can make claims and you know unlike unlike open it's much more sort of linked to the real world they have a more complex dow structure where it's connected to a real world company and so you know it seems like a little bit more of a bringing the existing system but recreating it into a deep pot into a decentralized manner yeah so this was great fun I mean I remember speaking with you and I mentioned it was sort of the episode around four years ago I think when he was just starting to work on Nexus

mutual and now and I thought you know this was interesting and it's amazing to see how far they've come and yeah I thought this was really cool especially the idea that you know we're gonna speak about this example as way as well right that you can go and compound and you can lend out let's say some you know Fiat stable coin and then you can buy insurance for for it on a nexus mutual you know for the case that there's some bug and the contract loses its value so this is really cool use case and yeah I'm impressed I think this is like one of the coolest dabs that I've covered in an episode episode

so uh yeah I hope you guys enjoy the episode we're here today beef you carp he is the founder of Nexus mutual and also board member of its are we gonna speak a bit about kind of the structure of Nexus which I mean excited about this I remember vaguely you know having spoken with you hang around four years ago about Nexus mutual and it was like at the very very beginning it was really exciting today to catch up on all the amazing progress that has happened and just soso kudos the website in the documentation are just really well-written and very very clear so well done thank you we've been at it for a while video it's great to be out there

and actually providing these to rip two people in the defy space especially right now so yeah really great to be here so you have a long background in insurance tell us a bit about like what was your insurance work like and how did your interests in blockchain come about I'm an actuary by training and I studied actuarial science at a university in in Australia and yeah what qualified as an actuary and I went for 10 years in life companies in Australia doing various roles pricing and reserving and product development and a bunch of different things and get more and more into a marketing side a bit later on when I was was there including them then switching to doing M&A deals and a bunch of

different stuff 2012 and moved to London and joined mimicry which is one of the guess one of the largest Rangerettes companies in the world and ended up working in their life division in the UK and did a few different roles but ended up being the CFO for their life operation so kind of ran a team of about 30 people actuaries accountants analysts and stuff like that and did all manner of things Capital arbitrage shifting money around balance sheets and doing all that stuff kind of really at the back end as a insurance value change so that was kind of really interesting for a while but I did get a little jaded at the end because it was kind of it was a bit far away from like the real impact on people and like actually it felt a bit too

abstract to me sometimes I guess that's kind of the insurance II stuff for quick wrister on that but I I'm the kind of blockchain piece I you know stumbled across Bitcoin about 2010 I think and you know just kind of went down the rabbit hole started playing around I do remember I had absolutely no idea what I was doing and you know it was kind of mining on a web-based miner I'm doing some really really basic stuff storing private keys on public internet websites and stuff it was just I remember doing like all the things you shouldn't be doing play around with for a while I found it really fascinating that especially you could send money from one person to another without anyone else in the middle that's just kind of really

blew my mind a bit and then I guess I didn't really know what to do with it because I guess payments wasn't really my thing I found the tech fascinating ago didn't know didn't really know but I could personally do with it so I put it down and honestly I didn't really come back to it for quite a while I think it was kind of late 2015 or early 2016 that I kind of heard about theory and then kind of started getting back into things and investigating and stuff and I think that's kind of when it really kind of clicked for me because you know if you can write an offender statement then you can write an insurance contract I kind of like really kind of started doing research and stuff then and know that kind of the early stages of I guess Nexus and the idea formation and stuff

so yeah that's kind of the brief history so back when you were at some of the you know Morris legacy insurance companies where they at all interested in blockchain stuff or work even if the company itself wasn't were there like murmurings amongst individuals or other people really interested in this or not really I think is mainly only talked about our winners of Munich hurry so I can only really comment from that point of view but it was definitely a topic of conversation you know it's new tech it's on the road you know on the kind of road map of things to look into and they always go through you know scanning stuff on the horizon and things so that they look into it I guess blockchain was kind of always lower down the priority list and other

things that were there are more kind of specific insurance stuff so like IOT or AI or machine learning type stuff is kind of always better higher up the list in terms of priorities I think block don't got put in the bucket of this is a potentially useful tech especially in the insurance world if things are very interconnected but it was more in the bucket of this could save operational expenses and there are a lot of potential technology solutions that can do that so it would be competing against a whole bunch of different things so it never kind of really got the high that will focus there were a couple of us that were always talking about it because we were kind of deep in the rabbit hole maybe a bit more on the crypto side but it didn't tend to really get much traction within the in the companies

at the time I was working was called me Mon acts which was started doing kind of enterprise Hiram's they start in 2014 but I was there like 15 and 16 and that was you know significant part of the interest was from insurance company and you know there's like pilots and lots of conversations of insurance company including Munich Re right there's a Marcus there who was later there were social insurance consortium spun up we spent quite a lot of time back then and and they seem to be many also doing various like pilots and projects maybe from the outside it looked like they were quite interested maybe even most of them banks from our vantage point but

maybe if any inside it look different but I'm curious how has this all progressed like what has come of these things or any things here going like life in production being used today or is this kind of died off there are a few that are alive today I'm working I think it's fair to say that everyone was like very interesting because there's a new tech and it can do something fundamentally new so they all wanted to get in and make sure they could understand and perhaps cynically it's so they could answer to their bosses with when the question comes and they'll want to run a pilot to really kind of get into the details but I think it then the step of taking it from a pilot to the next stage of opening actual implementation in a real case that's a there's a massive drop-off

there I think there's a few that are actually live so p3i are doing a bit of a live product now where they're essentially running this consortium amongst the different ventures and insurers where you place like access of loss catastrophe insurance on chain and I mean it actually works quite well because you can coordinate the placement there's like you know if you want to take a huge risk you kind of slice it dice it between different entities in the insurance world and coordinating all the paperwork and who's on for what risk is actually a large part of the problem and so blockchain is quite good at coming notarizing that stuff and making sure that sign offs are all done and you know you're not chipping paper around and all that all the time so that's working quite well I think there's also a mask

one which is I'm kind of shipping supply chain type stuff which is an interesting problem because when the ship moves into different waters and different jurisdictions different people are on risk and so you have to kind of know exactly where the ship is and like you know tags and stuff like that so it's basically about data integrity and those such things which amongst multiple parties so you know book chain works in that context a lot of like what b3i and stuff is doing is sort of helping coordinate these like existing insurance companies and you know like you said making the operational cost lower but then you took a sort of very different route where you know you went a very crypto native crypto first like rebuilding insurance from scratch maybe

all of your colleagues who were also interested in crypto went down that's very more private blockchains approach what drove you to building like public decentralized on aetherium to me it was always a very natural feat I guess like basically insurance is a coordination problem and it's an intangible good it's like you can actually have it running on a blockchain very natively and it works really well because essentially block chains are good at coordinating people and it's especially incentive mechanisms and that because it's no there's no physical but necessarily physical good it's it really kind of fits natively like a currency in some ways so there was always this very natural fit and I also saw that there was like massive inefficiencies in the current system about moving capital around and also

just paper administration and stuff and so I saw that there was a big opportunity there and I guess the other thing was I also noticed that all insurance companies really do custody money on behalf of people so essentially you pay your premium and the insurance companies look after it and hand it back as claims and at some point and as long as they do that appropriately then everything's okay and so you've kind of got all of these rules have like regulation and capital standards and all this stuff to make sure that they've got enough money and they do the right things I thought that that could be done in a programmatic fashion using smart contracts and then all of a sudden you don't need all of these surrounding stuff did the insurance companies not also do like sort of the risk analysis

and all that stuff or is that sort of outsourced to other entities it's a combination but yeah they do to do that risk analysis and stuff as well but I guess what I'm saying is that you need certain inputs like that and you'll need them in any way shape or form but I just saw that there was massive inefficiencies in the kind of agency risk I mean that's not all reason actually the regulations exist is so that you make sure the insurance company does the right thing with the money and it's got enough to pay claims when they come that's basically where we're thinking that this could be a whole lot more efficient and also insurance is kind of historically has always kind of come from Mutual's like which kind of community or people owned insurance companies rather than necessarily

shareholder owned insurance companies and so they kind of work for the I'm it's a very natural fit like you've got a community that shares risk together if things go well that they should all kind of benefit and they step in when times are bad and so it's a very community native solution and so that's kind of one of the reasons we wanted to do it was to kind of make a scalable mutual that everyone could join and scale the capital of which is kind of one of the problems that Mitchell's have but anyway so it's kind of a it's very like native watch-chain use case even in my opinion it's always been talked about and it's hard to do and I guess yeah wanted to give it a crack even though we've got the expertise you mentioned Mutual now so my understanding is that there's

different models of insurance in this you know mutual or mutual insurance is one of them can you explain like what mutual insurance is and how it differs from maybe the more standard insurance model yeah I mean there's kind of two general models once the mutual one's the shareholder owned essentially it just depends who owns the profits that come out of it really so a mutual basically if it if it makes a surplus or profit makes a surplus at the end of the period end of the year then it essentially gives the money back to the members in some way shape or form whether that's very discounted premiums for next year or past repaid or something like that that's how kind of how they work and so there's always this view that we're

working for the members and then obviously share how the companies Germany goes to pay you the shoulders dividends et cetera which is more common in the real world today the shareholder companies largely dominate there are a few big Mutual's but they tend to mostly what happens is they may just tend to get swallowed up by the shareholder companies and that's primarily because of capital efficiency so initials can actually only raise money from their members and if they're members of the community then they're not actually know massively deep-pocketed capital markets like that that's kind of the reason you would shift to a shareholder based company the e can access more flexible capital essentially

that has made that a lot of the Mutual's kind of die neutralized so they what they do is they turn into a shareholder company and give the current members shares in the new shareholder company there's kind of a proxy for their ownership and that's kind of how it works so Mutual's have always have this structure or problem scaling up into big sizes and that's kind of one of the reasons that design Nexus the way we have trying to use token economics to assist in that scaling problem and our theory and our goal is to kind of break that scaling issue that Mutual's have and be able to get big I didn't totally understand why this scaling problem exists for mutual companies can you explain that a little bit more simple example if you've got ten people in a mutual and they each put

a thousand dollars in then you've got ten thousand dollars and so the most cover that you can provide is really like up to about ten thousand or you know because you have to have enough money to pay claims if you get a hundred people also in the same thing that's fine but to grow fast you need to grow the actual you need some money as capital to actually cover bigger and bigger policies to start with and so one of the problems that you Mutual's have is because they their members are usually just regular people they don't don't like have a hundred million sitting in the bank then they you have to essentially put limits and they can't grow as fast so shareholder based company is essentially what they do is they just sell equity and just get a

massive capital injection to kind of kickstart the process and then they can sell really large policies and wrap up quite quickly and insurance is really a scale game it's all about capital efficiency usually the it's all about it it's a huge leverage play so for example you know like banks are kind of leveraged right insurance companies are even more leveraged so they sell like many many many more times cover than they actually have capital to pay so like basically if everyone claimed at once they would never have enough money but the whole idea is that their likelihood of everyone claiming it once it's so so low that it's okay for example like AIG covers about 60 times more amount of risk that it does actually have capital on its books to pay and so that's kind of a whole idea

and that ratio that number becomes bigger and bigger the more diverse and bigger you are and it's a big scale game and so what happens is the Mutual's go they have a good community product and service and all the rest of it and then they grow to a certain point and they need some extra money to grow bigger and they're usually the easiest way of accessing that is get bought out by a shareholder owned insurance company or convert to a shareholder one themselves it's kind of like the back end of the insurance capital markets really that's kind of driving this stuff so given all of that why is it next is mutual then if we seem to see that Mutual's have these sort of scaling problems why did you sort of design in at a mutual model

the whole idea is that meteors have this going problem because they can't get in like they don't have any equity of the cell so they can't get in equity but what you can do is actually you can tokenize the rights of the mutual and while that's not actually getting in equity it's actually getting in a more flexible capital structure so you can have access to the same flexibility of the capital but without going outside the constraints of what a mutual is I mean talking only raise money from the members but you can still have a member that provides funds to the mutual and your tokens in return but they're still a kind of a member and so you're kind of accessing the same type of capital or similar type of capital but in a different way and a blockchain native

solution can actually make that happen really easily and so especially we can use tokens to give us a capital flexibility where Mutual's don't have it right now are there like some sort of regulatory benefits to being a mutual and is that why like you kind of structure it as a mutual with like elements of shareholder based this kind of goes into a question I was like do you run a fully decentralized EFI protocol versus what we're doing which is kind of is hybrid which is effectively a Dao but it's linked to a legal entity in the UK and the primary reason we did that was because it gave us regulatory clarity because insurance is like one of the most heavily regulated industries in the world and so we wanted to be able to get some clarity

on that side of things and there's what's called a discretionary mutual structure in the UK and that's specifically not regulated as insurance business and so people use this today to sell neutrals I've got like regular Mutual's then and if X servicemen's and Mutual's retail shop owners Mutual's like there's like you know build real stuff and so we're basically just using that structure and automating it fire smart contracts that gives us regulatory clarity I guess the other element is that we knew we couldn't build this in absolutely fully decentralized way from day one and so we couldn't like go the kind of full decentralized route no regulatory you don't care about the

regulations at all from day one so we decided to go with this hybrid model which gave us the clarity let's talk a little bit about what's the big vision for a nexus mutual where do you want to see this project a decade from now and when you think of the benefits that do you think Nexus mutual or maybe insurance solutions based on blockchain bringing in general what are the most the biggest pros while we're aiming for is guess a decentralized version of Lloyds of London anyone should be able to bring any risk they want covered all that quirky stuff from any region in the world kind of in a permissionless basis and they bring it to the mutual and if the mutual wants to cover it they can and then the person can pie cover a kind of

route to that is more more kind of stay within crypto because it's kind of fits with the nature like that our early users are gonna be we need providing products for early users actually early adopters and then move outside that as we grow in scale there'll be a lot of luggage journey but that's the kind of goal scale into a whole bunch of different things and they're a whole bunch of underserved markets out there like Mutual's work really well where you've got we're not gonna work like we thought Oh cover in the US that's like a heavily saturated market highly competitive and stuff like that muchos work well where you've got an underserved market with a specific need that's either niche or they can't get access but some other reason the price too high or whatever it is and so Mutual's work really well in that

situation and we think we can access those types of benefits give access to those types of communities really quite easily and efficiently talk about why is insurance so important like especially in the defy world I was riding with a friend once and he gave this like really great explanation to me about how insurance is trust lessness where like you could go to traditional company and say their systems are 99.9 percent secure and we go pitch like blockchain to them over like look using our stuff you're a 99.999% secure or they're like well in the centralized system that point one percent of the

time we're insured so for us that's trustless your system is not do you think that's like actually a fair representation of it when it comes down to it like every industry in the world has insurance for basically the reason that well there's two real reasons one you don't necessarily have the capital yourself if something goes wrong but you say you're good self-insure right like so if something goes wrong and you've got the money sit in the bank then you're going to get back to anywhere then you could decide to do that that's fun but obviously that doesn't work for the vast majority of people like you can't just rebuild the house if they have a fire or whatever so that's kind of one reason the other reason is like the marginal cost of getting from being 99 percent secure to 100 percent secure is really really high and so you're

better off offloading that risk to a insurance company as I guess like natively like the same conceptual something applies in defy where everyone's is really doing something I'm really so black on bikes my contract security as an example and you know different teams can afford different audits and verification and all that tough stuff but you're never going to be able to get to 100% because the emotional cost of going from I'm really confident this doesn't have been - I'm absolutely sure that it doesn't is like just so astronomically high and so it's much cheaper to actually get essentially and assure an insurance product for that it's been really interesting to like understand a bit how important insurance has been historically like the same friend he was actually explaining to me how like a lot of maritime law was

actually sort of devised by Lloyds of London as a way of setting the rules to like protect its own you know anyone who wants to get insurance from Lloyds has to like abide by this the laws that it sets for the sea and eventually those got canonized eyes like the maritime law soups is a good example right like a lot of those ships like the merchants wouldn't put their goods on the ships unless it was insured and the people who are insuring it effectively said you have to make these standards as wise I'm love insuring it and so it's kind of this standards type thing and then once that happens they it actually just Spurs massive economic activity because everyone's confident in things and they're protected with the downside so you know you're much more willing to

take risk if you know that the downside is protected you know you've like shooting a spacing rocket into space like that doesn't happen without insurance like no one's going to give you permission to launch a rocket unless you have finished as policy advice so so you know it just applies any kind of every single industry and so I guess our faces is like you know if defies building uh-oh and theorems building this parallel financial ecosystem then we need some insurance to kind of make it work could you walk us through a little bit now about how does it work like what would be the process let's say I'm a user of a protocol let's say maker Dow right I have some CDP's and I want to insure against smart contract risks on maker doubt can you walk us through

what the user process would be it's relatively simple basically you go to power up and select make it down so it's like what you want cover against whether it's maker compound or unit swap or whatever and then you choose how much cover you want so for example if you've looked up 100 eath in your CDP then you may be 100 eath worth of cover if something goes wrong and then you choose the time period you want cover for three months or a year now you get a quote it comes about if you're happy with it then you purchase it pay the premium you also have to sign up as a member of the mutual which does involve a kyc process on outside because you join the legal company in the UK but that's basically it and then say an event happens you know there's a hack or whatever then to

submit a claim you basically just submit the claim one transaction and then that triggers a an item essentially a workflow item for our claims Assessors to vote against so there's a we specifically have a voting process here rather than a programmatic outcome and the reason we kind of did that was it gives us heaps of flexibility in terms of what we can offer on what products we can offer essentially what that means is the claim is this or as in aggregate they're the article they're bringing in the information and they're they're making that assessment on average and so assuming it gets paid then you automatically get paid your claim to your address and that's kind of the complete user flow so who's doing this voting is it just the members of the mutual or members of the mutual so any member of the mutual can decide if they

want to optionally vote in the claims assessment and to do that they have to take some of that Native token to do that and that's kind of their voting weight and then it's also their potential looks like a bond so if they don't maliciously they can lose some of their bond that's kind of the general process of how it works how does this native token work let's say I joined the mutual do I get sort of awarded a fixed amount and is there a way to get more Native cocaine if you want to be a larger mutual holder larger members it's basically we actually run a bonding curve so we're actually one of the first live bonding curve projects out there which is interesting but the first step is you sign up as a member pay this with very small membership fee and then you

can decide to purchase an exam for ether so you put ether into the mutual and based on the bonding curve price at the time then you get an exam and you can always catch this an X amount so if you wish there's a few things you can do is kind of like an exam holder so you can vote in the claims assessment by staking that token against does your bond and you can also get involved in what we've got risk assessment which is like the pricing side of things so you can for example stake on make or a compound like basically you can pick this the protocols that we think are secure and then you earn rewards when people buy cover and you could potentially get burned if there's a claim so there's kind of a plus or minus there and you can also click voting governance actions

as well those are kind of the three kind of expert level stuff that you can kind of come in and so you can buy more tokens free through as you wish so one thing I found interesting here is that you have these actions like you know you mentioned the finger of luck okay you can stake tokens on a particular protocol to sell small contracts to say like oh I think this is secure and basically you're putting up some sort of collateral to back that and you make money based on based on the amount of collateral you know but at the same time like you have like various investors who are you know these VC funds who like they may be knowledgeable about like you

know Krypton a high level and stuff but not about smart contract security so I mean it seems like the natural thing here would be to have some sort of you know delegation function right where they're people who like specialize on risk assessment other others on some other function others that have capital and there's some sort of you know basically deal between them similar to you know proof of stake with like validation token holders is that something you're considering or does that make sense to you no I mean it does make complete sense and it is something we're considering I think um we're probably the first steps are going to be we're actually releasing a new staking mechanism very soon and the main aim of

that is to make the process a bit simpler and also enhance the rewards for the staking so kind of we focused on that to start with but then we're going to be looking at things like like step one might be you know allowing you to easily follow someone else so just kind of copy someone else's staking process but then also you know enhance the system so you can specifically delegate if you wish so yeah I mean that makes it doesn't make complete sense to do that it's very easy to just like sort of copy what someone else is doing let's say I just see what you know you know opens up and I just go ahead and just automatically copy whatever they do do you think that like they might be getting sort of under rewarded for the

amount of work that they're doing in the system like they're the ones sort of doing a lot of the actual like I guess I'm putting I'm putting a lot of cap but how do we make sure that they are sort of appropriately rewarded for the amount of contribution that they do yeah I mean you're always going to be able to follow someone else say like we can't actually prevent prevent that so I think the important point is that especially we like with people like I've been selling this stuff like I mean if they want to get more involved they can earn more rewards you know they it's up to them to kind of decide how much more money they want to put it risk if they're happy - happy to back themselves a lot and then they can get more but with a delegation maybe we can have like

a yeah a fee aspect to it as well but you know maybe the the person who delegates keeps owning 90% of the rewards and 10% go to the person they're delegating to or you know something something like that so if you if you want to tell you to someone like they maybe have a fee on top of sake so there's definitely models around that that we can that were potentially looking at to make that process easier so yeah could you maybe talk a little bit about how the how does this whole price Discovery Park work because obviously this is one of the most complex things and any insurance system and you know we had open on last week

and even for them you know how to do the initial price discovery is what I'm always on the hardest part so we walk through a bit about how that works in Texas Mutual yeah so um I guess another thing that's soon to be changing very soon but but essentially if we boil it down it's basically staking so this is going to be a formula behind risk and the more that's fake to another position specific risk the lower the price so it's kind of like you conceptually it's like using a prediction market but it's a bit more structured the essentially you just kind of one measure and and the not very much dates obviously high price and then it comes down I mean a bit of a curve so a the more and more stakes you get the lower the price with minimums

and stuff what ifs takers doing here they are staking on the security of particular contracts yeah that's correct and so if let's say there is a bug in that contract the stagers there MXN gets sort of seized first before the claims are made out from the main pool is that sort of the main risks that the stickers are taking on yeah basically so what happens is it's like an ik that there's one capital pool where all of the funds live there basically anyone who wants to purchase an accessory for anyone and whenever someone pays a cover amount and it goes into the common pool and claims go out of that pool as well and also investment

earnings on the pool just stay in the pool so that's kind of conceptually there's just one pool of funds and that's kind of how insurance entities like work they kind of got aggregate the capital all together so that's just kind of the first concept and then over the top of that is like the incentive layer which is the staking and the reward mechanisms and stuff and so when a claim is paid the the ether or dye which is bought the covering gets actually paid from the capital pool and concurrently the stateís would get burned up to the claim amount if there's enough stake sometimes they may not be enough stake so it may not be fully covered by stake but the kind of mutual would be taking the risk by itself then but essentially if there are if there is taking it when

you say burnt I mean you have this capital pool where things kind of accumulate and then you know out of this let's say MXN is paid to cover claim so when you say oh the stagers now get burned basically means there's not like literally Tobin's being burned but that there is some part of the capital pool that you know allocated to those takers before and now it basically it's kind of appropriated so you kind of I think the whole idea is that you've got like one one to one black token a whole bunch of token sitting over the top of the

capital pool and they kind of have shared ownership of their capital pool because they're like the members their membership rights right and so whenever someone puts money into the mutual they get more membership right so it's good to come on the bonding curve so it's kind of step one and then and then what you can do after that is kind of the staking so perhaps conceptually it's easier to kind of just stop there for the moment and if you're just a holding an exam for example then you're a member of the mutual and to the extent that the mutual makes a surplus like so essentially the cover prices exceed the claim payments and the investment earnings of the capital pool then you get you get your kind of part of that so you get to kind

of guess sharing that side of things then as an additional optional step you can increase your risk reward position if you'd like to by staking and if you stake then when someone else purchases cover you earn additional nxm tokens and if there's a potential claim if there's a claim then you can potentially get burned I guess you can take different risk reward positions depending on what your appetite isn't and what type of memory you'd like to be okay that makes sense I mean one of the things when I was reading through your you know explanations on the website that stood out to me and I found it a little bit strange maybe you can explain it so if I'm buying I'm buying some insurance coverage right and let's say I'm paying

a hundred MXN for for this insurance coverage then the way it's said in the in the documentation is that ninety MXN of those get burned and ten get kind of allocated for a future you know if I want to make a claim I have to put up a little bit of deposit at the same time you say there was 20% of the cover paid which gets paid out to the stagers so it mean it seems does does that mean you're like burning MXN and then at the same time you like minting let's say in this case we have a hundred so you're burning 90 and you're minting 20 to pay out of

those takers yeah yeah so I mean rather than like kind of its easiest to kind of look through and look at the net result right after after after everything right and so the whole idea here is that basically from a user point of view you actually buy cover in either ether or die and then what happens is step one this happens all in one to one transaction step one you buy you put say you're buying coverin ether it converts it to nxm and then it burns 90% of that makes sense because it buys it on the bond again yeah it destroys it immediately right so it's kind of a side point so really the ether goes into this in the pool and then it would be like let's say I can say and I just want to buy an

xn I don't want to cover so I'm putting ether in the pool and I'm getting a mix n from this bonding curve whereas you know sunny he doesn't care about him xn he just wants to surance cover so he's putting ether in and pool the same thing happens am I saying is taken out but then he doesn't get him xn but that gets burned or / nor any be saying gets burned 20 percent quasi gets to some stake for that contract and 10 percent he kind of keeps for you know to take future actions around his insurance yeah that's right yeah cool yeah so and so I mean I guess it's important like all of these kind of metrics and stuff we're kind of we're basing them on like insurance company economics so that's

kind of a whole idea being that we we priced the cover at the right levels you know obviously we don't know what the claims are going to be right now we need to get more experience and stuff like that but the conceptually the idea is that leaves a little bit of surface in the pool and also for that 10% for the cover purchases that's kind of ideally it's kind of like the mutual like you can catch that out at the end if you'd like to or you can use it to buy discounted cover next time around so it's kind of like your ongoing participation in a mutual so it's conceptually that's it's that's how Mitchell's work so going back to that question about the sort of price discovery so alongside the staking there's like other things that are taken into account as well I read about like

you know sort of how long a contract has existed or could you talk about some what are some of the other factors that also could sorry yeah so I mean that's very interesting so we've it basically is kind of trying to proxy how battle-tested the contract is how much it's been used how much funds is of God how long's have been on main and how many transactions that have and so that was um that's kind of that that initial pricing algorithm is kind of a first attempt at kind of blending old insurance world with staking mechanisms and stuff we if we've worked a lot on on these pricing stuff over the past I guess we've been on Maine in about a year now I'm so got a little bit of experience on stuff we haven't got many claims we pay down the B's at X 1 but that's that's it

so far but what we've learned is that that algorithm doesn't work the best in the situations in all situations and so what we're going to do is actually simplify that down and just bring it back down to staking and so what we're trying to do here is like bring the minimum amount of information on chain and allow people like if someone's got like a really awesome model to assist small contractors they can do that and they can run that off chain they can do whatever they want and then they can essentially condense that information into whatever they want to stake and then they just put that on chain and so the idea being that we don't care what tools you use or how you can actually come up with your view but we just care how much money you put behind it and so

that's going to work for the smart contract cover but that pricing mechanism can actually work for anything because it doesn't actually reference anything about the particular risk just reference to the state so if you want to like offer earthquake cover or a hurricane cover or like in anything you can you can effectively do it using the same mechanism and so that's kind of one of the real goals that we're trying to get to longer term how do you account for correlation than here right now you don't cover Oracle risk but let's you know let's say eventually you do a lot of stuff does Kate like depend on for example the maker Oracle's particularly and that's like this point of correlation so does accounting for the higher you know maybe because a

correlation you maybe need higher premiums for contracts that do have like maker that touch the maker Oracle's is this something that would also be accounted for within the staking system where is a additional functionality that's needed to do that no so I mean generally the best way of managing correlation risks through capacity constraints rather than through pricing just in general with insurance that's kind of how it works you don't you don't want to like basically you know if you take you know you're exposed to like business insurance in New Orleans like you know if I were like a whole bunch of you Freni translate if a hurricane hits they're all gonna go at the same time so you basically limit your total capacity

to that type of event that's the kind of best way of dealing with it rather than through specific granular pricing so that's kind of the way we're headed as well to be honest it's we're still kind of working out the best ways of actually capping and managing those correlations and we've definitely got some few ideas but that's definitely really kind of higher-level expert stuff that we're trying to work out the best most efficient ways to get any information on chain but it's more likely to be managed by a hot on we're gonna make sure that these five risks are grouped together and their total capacity is not greater than X so that they can move in between that but as long as we kind of kept the total so that the mutual doesn't go bust if there's like one event that hits five things so we have some limits now that

are a bit more crude but we're going to be looking into ways to manage that a bit better in the future and right now you do only the smart contract risk what are sort of the plans to expand here and what and what kind of happens when things get the line sort of get blurred so you know Oracle's aren't a good example of it's sort of blurring the lines or I just read just last night that there is like some bug found in the Argent wallet an argent is sort of this like interesting thing where it's like half of the stuff is being done on using smart contracts but then a lot of the stuff also comes from like you know their app and like you know would you also cover like it's often on clients need to be smart in order to know how to

interact properly with smart contract if the clients have a bug in this incorrectly with smart contracts does that count as something that's covered by Nexus mutual right now not right now so we definitely we focused down onto the essentially solidity cover is probably a better actual description of what we're offering so technical risk and the solidity contracts I'm so you know if you if you send the wrong information to the particular contract because the UI had a bug then that's the kind of thing and same ends I guess same with the Oracles right now having said that we we started there because we felt it was a kind of contained thing that we could like at least get some handle on I mean it's completely new risk any way that no one else writing but we thought we'd start there

but the plan is definitely to expand because you know if you're interacting with defy you basically want comprehensive cover and say in case something goes wrong and so let's we know that the customers and members want that it's just about working at how we actually get there in a good way that we can actually define the terms of when a claim is valid and when it's not and how you prize for that and stuff like that so we've got some good ideas on that and and essentially build on the architecture that we've got and we're looking to do them in the near future as well what are the volumes you have so far like let's say if you take something like compound or like Oh what are the most popular insurance policies you have

and how much insurance is so far been taken out so we've written million up to 99 million worth of cover so far we've got about four and a half million outstanding right now and the big ones are surprise surprise compound of a balance for anything with yield farming right now this is a beak is a big one curves also really popular and so it's basically the the really solid use case for us right now is like you can earn kneeled in defy via whatever method and then you can pay a lower amount and get some cover and then you kind of protect it against a large chunk of the Ring that's that's that's the big guys cool and so like what's the amount you have to pay to get like let's say i mean i

don't know and i guess i'm comp our most most rates are kind of low but maybe died or something like I think the tether I think the stable coins are kind of high so like let's say I would want to buy a year's worth insurance they're like how much would that cost yeah so right now the rates on compound unless the other big ones that one is 1.3 percent per annum so that's works quite well when you're earning 5 6 7 or 10 or whatever it obviously depends when the DSR went to zero and that was kind of the if that's kind of the risk-free rate not close to it then everything else kind of followed we weren't selling as much cover over the last three to six months but recently with you know tether being able to get a temper so

done on compound and stuff and runs I'm really really interested in this stuff again so I yeah that's that's we kind of we kind of go it along with the weed defy for the ride how do you deal with sort of like perverse incentives stuff that like you know you don't cover or career so there's no issue with that but like what about like you know let's say I'm gonna tack her and I found a bug in a spark contract I can sort of double my payout by first taking out insurance and then doing an attack right how do you is this something you do need to cover or there isn't right now so you can't essentially do it like a naked cover if you want to describe it like that so it's kind of like a credit default swap or something so yeah look this it does

lead to potentially issues we we started that way because it was simple to do I think you know what we want to do is actually have some level of proving of loss which effectively takes away most of that kind of attack type issue there however I mean the risk is still out there and like but if there is a bug there is a bug I guess there's a soft control we do have kyc so right now we don't want to kind of rely on that as a um prevention again it gets the tax but it does it's a soft one against hackers so yeah look I think with all things insurance you can never do this stuff a hundred percent there's always going to be a way of potentially defrauding the

system and the the way you what you have to do as a kind of provider is get that down to a costly enough level or a small enough level so that it still works as a whole because obviously if it's too high the whole thing whole system breaks down yeah look we're definitely still learning on that front and we're adjusting economics incentives and those types of things to to gain experience and things like that so yeah I'm not saying we've got this completely hundred cents off right now we're definitely trading could you talk a little bit about how this sort of compares to other models of insurance on in the defy world so prediction markets which in my opinion this actually seems sort of very

similar to prediction markets and in many ways and then the other would be something more more like open who we just had on the show last week where they do more of a financial eyes de proach to it yeah so I mean this essentially comes down like derivatives or prediction markets I kind of put them in those same a similar type of bucket and insurance pools like there's always going to be overlap big but they are actually distinct products in my mind and it comes actually comes back to the capital models and how it how the capital set up for example if on open and on prediction markets they have to be fully collateralized you know open they have to be more than fully collateralized which means if you sell a thousand dollars worth of cavity you

have to have on open one hundred and one thousand three hundred or something at home I remember that collateralization ratios and on a prediction market you have to have a thousand like in the pot so that if the event happens you've definitely got enough money to pay you know essentially the premium goes from the top person who wants cover to the capital provider and if you're writing risks that are quite likely to happen like I don't know ten to twenty percent likely to happen then that's a reasonable premium to go to a capital provider if you're talking about extreme risks that are very unlikely to happen so you say like a 1% type event if you're a capital provider backing out 1% event 1 to 1 then the most you can earn is 1% and so it's very hard to attract

capital on a fully collateralized basis for that type of risk so what happens is insurance pools effectively allow you to under clash wise they allow you to gear up so Nexus can at least gear six times right now and it can we're having to give more so we're currently riding at 120 hundred thirty percent so we're under collateralized right now which sounds bad but it's actually kind of the whole point of what we're doing and so that means it's more capital efficient so what insurance bills work really well for kind of extreme events that you need to diversify the risk across that's the whole point you know my capital pools covering lots of different options and like prediction markets and stuff they

work better when you've got like more at the money type risk so like open in Suffolk getting some really great success on like the F put options and stuff because their likelihood of something happening is a lot higher than 1% is you know that writing put options are close to the money or whatever so that's basically what it comes down to um they're very related products but it all comes back to that capital efficiency point so two months ago two or three months ago you guys did make out your first actual payment from Texas Mutual to walk us through the whole story of how it worked in practice and you know those this whole thing where like the there was first time claims made and it was rejected and then later it was accepted you walk us through why

and what happened there was a really interesting time for us and it was I guess really great that we ended up paying claims because that's kind of the whole point of what we're trying to do but there so the bzx hat hats or events whatever you want to call them happened there were two different events and there were two days apart so one that was on a Friday and one was on a Sunday and so what happened was essentially both of them were quite complex transactions involving a flash loan setup and then manipulating in on choice on chain price Oracle and then profiting from it at the end and so on Friday the first event happened within 12 hours or something someone submitted a claim and

and based on the information available at the time it looked like it was basically a clever arbitrage trade using an Oracle manipulation and stuff which and it appeared as though the contracts operated as intended and there was nothing wrong with them and so because of that the mutual members declined the client no there was no bug in the solidity code however then I don't know another like within kind of 12 24 or 36 hours after the event that B said X team put out the details of that kind of postmortem and it highlighted that there was a smart contact trigger that they weren't that they intended to actually prevent an underclass wise loan resulting after all of this but it didn't fire and so effectively it was

bypassed and then because of that the the claimant resubmitted so you can submit your claim twice and because of this new information available I'm sucesses approved it that's basically what happened it was just really a timing issue based on the information available when the when the events happened one of the first things we did was basically say to everyone that has cover please don't submit claims for another day or so so that you know we have some information and people have done some analysis on this type of stuff so yeah I mean overall we were really happy that kind of retested all kind of elements of the system and I guess one other interesting aspect how quickly our claims decided upon like you know were they you sort of have to have

been acted upon within 24 hours like why why couldn't we like wait a week before deciding whether or not to approve or reject the claim yeah there's there's some there's some time limits and there's like a minimum of at the time we've made these longer now but at the time there was a minimum twenty twelve hour voting and a maximum forty-eight hour voting and basically but once you get it like a certain voting power it just closes automatically after twelve hours so that's kind of what happened so yeah we extended those time periods to kind of handle this type of thing I guess the other other interesting aspect of like we have we've had like quite a few claims votes now that but all of them have been like over 99% of people voting in the one direction like so basically this means like a very very

high level of consensus about what the what the result was even if the discord channel is flooded by lots of people scrape screaming at each other which is which is interesting in the north south cool this is really awesome and I'm super excited to see how this is gonna play out now maybe we can just spend you know two or three minutes on this topic but I mean I know you want to go beyond smart contract covers also to cover you know maybe real-world risks what do you think are the most promising you know sort of types of insurance coverage that you can write and what what's gonna be the biggest change you have to go through to go from you know like pure the blockchain to you know this hybrid

blockchain physical world scenario yeah so I think one thing that's kind of critical is I make stuff work right now especially is the fact that to assess claims you actually need some data source or information to assess them and so that kind of naturally pulls you into risks initially risks like earthquake or hurricane or something where you've got solid data coming from mobile data sources about where events happened uh stuff like that and crop insurance and uh satellite data and stuff like that so that's that those are the types of risks they're more likely the first non blockchain native stuff so the big the big things that have to happen for that to become a reality

one fear on off ramps and that is slick to kind of abstract away all of the blockchain stuff behind the scenes so that regular people can just use an app and click a few buttons and then it's done I think the whole idea of Nexus is that you get a distributor that creates an app for earthquake cover in Mexico or whatever and they are in the distribution there and then we make sus provides essentially the risk capital directly into the application and so they're actually buying cover with Nexus but Nexus isn't doing the distribution and customer facing stuff is done by partners that are local and understand their products and the markets and stuff like that that's that's where these

types of solutions tend to work very well so that's kind of kind of me that tech side and I guess the other the other thing is you also need a few experts in the mutual that are interested in doing that type of stuff so right now we've got a whole bunch of crypto experts that are interested in crypto related risks but to expand into different types of risks you're also going to need some community members that actually have knowledge about those types of things and so you kind of have to grow into those things slowly with your membership base let's just touch on a last topic which is you have this hybrid thing of the the legal entity and the DAO like what is the relationship between the two then

one of the same thing so that the legal entity is operated by spark contracts running on ethereum so that's kind of how how it works there you have legal documents that kind of describe the these are the rules and if there's any disagreement between how the thing operates just look at the smart contracts that's the final determinant and then when when you want to change things like you have to go through a governance vote and all of that's kind of prescribed in the legal documents as the sync that kind of synced up and basically when there's a governance vote that means the terms and conditions of the mutual are essentially updated and so it all kind of works in sync which is which is kind of cool and I mean there are a few of these more coming out but we would guess one of the one the early

wants to do it as well I guess one other interesting aspect is which you mentioned right at start is that they're actually four five board members of mixes and they are the legal directors of the company and they have some slightly higher powers within the system but there's that kind of link to the two kind of the traditional world I guess this allows you to hold sort of assets outside of on chain so code sort of the insurance pool also all some I don't know Swiss francs in in a bank account somewhere is that is that one of the benefits of this it could do and I guess that is a benefit but it wouldn't be referenced by the token price so we're

trying not to do that one one of the big benefits it has I guess is two real big benefits one it gives limited liability on a permanent basis so you've done it you're joining a legal company in the UK if the doubt is something wrong you are only limited your liability is limited to one pound that's kind of the structure of the membership organization rather than if you were joining it down without this legal wrapper then you're potentially liable a lot of personal places for anything the doubt does mean that could be quite an extreme view but you know that's that's one of the benefits it gets and going back to close to the beginning you mentioned that one of the things that insurance company do is like you know what regulation does is that you provide a lot of the capital requirements and stuff

is there something similar to FDIC but for insurance like essentially governments have deemed that bank accounts are like a vital enough part of like the economy that you know they provide insurance there's the government all the government's also provide insurance like reinsurance to these insurance companies mmm no not specifically but they do set they do have like government-sponsored pools for certain things that it's often hard to get coverage for for example is like terrorism-related pools or flood-related pools or something like that um where you know the insurance industry may not be happy taking it on so you kind of nationalize part of it or provide some subsidies or something but but in general like the government

there's also this concept of being a systemically important insurer which kind of means that you have to pay some extra fees to governments and stuff so that you know they recognize that if you if that insurance company does go down it's going to cause massive economic impacts so they do extra to prop it up I guess what's the head like water if you look at the next twelve months whether the main milestones you hope to achieve yeah I mean there's a lot of work coming out as I kind of mentioned pretty soon and I'm like pooled staking and pricing and kind of incentive tweaking and stuff like that and that's really important for us and then and then the two the two kind of biggest things that we want to focus on one scaling that the Mutual's capital

pool and growing so we can meet the demand that we've got because a lot of people want lots of cover which I quite get everyone covered right now but that's kind of that's kind of one thing and then also working on distribution and sales and integrations and things like that to make it easier to purchase color so yeah rather than kind of coming to our side to buy Terra we want the vast majority of people to buy cover off our site wherever they're interacting with Defy whether it's your agent or a zero know what whatever they're doing so that's that's the time thing that that we've been looking to focus on it yeah just kind of growing and growing along with d-file that's the timing plant cool well thanks so much for joining us here I mean it's really impressive where you've come and I think this is really a

super cool that is life and functional and them to be able to see where you're gonna take into the next few years coke knife thank you very much and great to be on and great to talk to you guys it doesn't end here there's much more of this conversation and you can hear it by signing up for epicenter premium as a premium subscriber you'll get access to a private RSS feed where you can hear the interview debrief which goes on for an extra 20 minutes you'll also get exclusive access to roundtable conversations with epicenter hosts and bonus content you won't hear anywhere else go to epicenter rocks sloshed premium to join the community and support the podcast [Music]