Episode 348 - Taking ZK IRL with Cursive

Transcript

[Music] welcome to zero knowledge I'm your host Anna Rose in this podcast we will be exploring the latest in zero knowledge research and the decentralized web as well as new paradigms that promise to change the way we interact and transact [Music] online this week Nico and I chat with Andrew and Vivic from cursive we catch up with them about what led them to start working on a series of activations and experiments using programmable cryptography we map out the timeline of these activations often they were part of in-person events including events like ZK Summit 11 and ZK Summit 12 last

year they walk us through some of the challenges and also insights they had as they speedran the deployment of experimental applications using experimental libraries with real users cive is one of the teams I feel are on the edge of the application development front in ZK and in the adjacent Fields like MPC and fhe and they've been working on some Concepts that I think could be quite powerful in the future Concepts like passively sharing digital signals from private user devices or what they call digital pheromones as well as the idea of narrowcasting Nico and I have actually mentioned these Concepts before on the show so it's great to get a chance to catch up with the team who invented them now before we kick off this episode I want to share a few things from our EOS system first if you haven't seen it yet ZK Summit 13 is

happening in Toronto on May 12th applications to attend just opened space is more limited than usual this time around as our venue caps out at around 400 so we are encouraging people who really want to join to get their spot secured early it is invite only and we do prioritize previous attendees and long-term members of the ZK community so if you're newer and you really want to join just aim to apply early for a better chance of getting a spot the application to speak can be found in the same form the deadline to apply to speak is March 15th next up on the ZK hack front if you don't know ZK hack is a hub for learning about ZK if you're just getting into the space this would be a really good place to check out one of our projects there the ZK whiteboard sessions just wrapped up we released our final two modules in the past month

these were hosted by Dan Bonet they were focused on Fry and proximity proofs we also completed our first study group that went through these videos kind of in detail but we are looking to launch a new one in the coming months focused just on these fry modules so if that's interesting for you be sure to head over to the ZK hack Discord I'll add the link to that in the show notes that's where we organize the study groups I also just want to mention that zik hack Soul has been unfortunately postponed we weren't able to secure the venue that we needed and so we've decided not to do the event this year in the meantime though we are hosting ZK hack meetups in sopia around the ZK proofs event and in Denver right around the eth Denver event we are also developing our next hackathon it will

probably be in Berlin in June around the protocol Berg and dapon events we'll be sharing more over on the zik hack channels in the leadup to that so I think that wraps up all the upcoming events hope to see you at some of these now here's our episode with Andrew and Vivic from [Music] cursive today Nico and I are here with Vivic and Andrew from cursive welcome both of you to the show excited to be here longtime ZK podcast Fanboy oh I'm so glad Nico thanks for joining hello hello this episode is a long time coming we've actually been trying to schedule it for the last two months or so Nico you and I we did an episode back in October where we talked a little bit about some of the cursive activations

and experiments and then in our end of year episode we also mentioned what you guys had been up to so yeah it's so good to have you on so can actually dig in to yeah all of these events that you've been present at and what you're thinking in terms of yeah using programmable cryptography in the real world so let's kick off with a little backstory on the two of you where did you get your start what got you excited about this and maybe at what point did your path start to to merge Vivic let's start with you yeah so back in 2016 was the first time I started doing cryptography work um I was part of this like uh research program at MIT uh where I was paired with some grad students to work on computer science projects and they just like randomly

slotted me into cryptography and yeah I like started out doing some stuff with consensus Alternatives like proof of space so very cryptoeconomic um but involved cryptography Merkel trees and then I did some authenticated dictionary work with alen tscu who I think has actually been a guest on the show as well oh yeah yep um so we worked together for two three years uh using kzg commitments for like authenticated dictionaries and then kind of just dropped off cryptography for a while because I kind of thought I don't want to be an academic and everything in Industry seems like it's weird crypto scams or something like it was like during like nft craze oh um and I like briefly interned at like a crypto trading firm and I was like this is degeneracy I don't want to do this

crypto is scary and weird it's true absolutely and had the exact same process when I yeah studied cryptography and then I was like I I can't get into all this crypto scam nonsense yeah like I have to work somewhere else no exactly and then I I did like machine learning research for a few years at MIT and then near the end of graduation one of my best friends aush who's also been a guest on the show um he's like good friends with Brian goo the founder of xerx Park um I think they go back to like high school okay and he told me they were spinning up the this new lab they're doing this like weird ZK thing and I was like what is this like you know it's probably another crypto scam and then I started reading the vitalic blog posts and I was like oh my God this

is like KD commitments like this is the same stuff I was doing it's just like applied in a different way and so yeah I basically started with them in like late 2021 early 2022 uh doing some initial experiments around Anonymous speech nice so that was zerx Park what were you I mean were you helping out did you join them as a like an employee or were you just kind of learning along were you kind of a a resident I actually don't know how it was structured back then early on I think it was structured very similar to PSC in that it was essentially like giving out a bunch of Grants um a lot of them were like Brian's ideas for various projects he wanted to see including the stuff that I was working on like Anonymous speech was something that he was championing for a while okay so I was just a grantee there

for about a year and then I was briefly a full-time employee working on sup pass M were you there during like the Dark Forest era or did you join after that no I was not I I like knew it was happening um I'd like known Brian from some other programs and I knew they were working on it but I was not involved with dark Forest okay and what about you Andrew how did you get your start I guess I had a less involved past with cryptography I never studied cryptography in college but but one of my friends from doing math together kind of reminded me that programmable cryptography existed he's basically like oh this ZK thing is really cool you should go check it out and I was completely fascinated this was

maybe in 2023 okay in springtime yeah so quite recent actually um and I got into it purely from the matth side I was like this is completely magical the idea that you can prove things without revealing any extra information is completely absurd and it makes no sense and the fact that it's just built on algebraic objects and numbers is incredible so that's kind of what drew me in at the beginning and after a little while I was like okay this is really cool I want to work on this for a little while let me go meet some people in the space first I joined a psse project for a few months it's working on identity wallets for a bit based on zero knowledge proofs and

then afterwards I went to meet a bunch of people who were working in ZK I met lockman from person I don't know if he's been on the show um but lockman and I had a conversation where we were talking about how ZK enables people to express themselves differently how it can be connected to Identity in the physical world uh one cool concept we toyed around with was the idea of being able to change an outfit or wear different uh pieces of clothing and basically become a different person have different identity personas and from that lockpin was like okay I know someone else who has similar thoughts or is thinking in similar directions and that's when he introduced me to Vivic and we've been working together since then oh cool Vivic you

were doing zoo pass you sort of just mentioned that this is zalo zath yep same one so I guess did you guys meet after that yeah so I was yeah working on the zass team and actually I think where cursive like properly started was at suu so um there I met Rachel who's like currently uh like one of the co-founders and like our head of design she had been working at PSC for a few years like helping with design for a bunch of projects there and I also met uh Alia who was like a core member of our team uh during the early stages uh like also kind of a design and like comms background and so like it was near the end of zulu where we all kind of met and built out kind of an initial NFC

experiment uh using like signatures and NFC cards which we can get more into in a bit and then I think that was like May 2023 and I met Andrew like July 2023 over video call I actually when I first called Andrew I thought he like thought what we were doing was really boring like I was like oh I don't think I'm ever going to talk to him again that's unfortunate um but then he reached out again I was like oh okay cool nice let's do something did you find it boring Andrew no no I thought it was awesome otherwise I probably would not have reached out again that's hilarious I I didn't know that cool you also had long hair I never actually saw long hair Andrew in person I just saw it over that one I did have long hair video there's very few people

in the ZK space who have seen that rare version before we dive into cursive because that's kind of what I want to cover primarily in this episode cursive and the experiments you've run I don't know if we've ever talked about zup pass on the show and it might be worth it to talk a little bit about what that was because yeah this is like atsu zalu that TW Monon pop-up City kind of the first one that prompted I mean since then there's been lots of variations and and I think there's new zalo in different places but yeah there was this zoo pass so maybe tell us a little bit about what that was yeah definitely so I was one of the first two devs on the team and I was there for sort of like roughly the founding of it um I would not call myself a Founder but like I was

basically just like hanging out with Brian around the time when he was coming up with the concept for zpass and he was like do you want to work on this and I was like yeah let's do it and yeah I mean I would say it's basically an identity wallet that I think its Innovation as far as I can understand is like they targeted a specific inperson event and also one that was very friendly to trying out new technologies yeah were able to sort of like digitize a bunch of various like tickets that were involved with events and like other kind of like data for citizens of susulu or other sort of like pop-up cities um yeah and I think in in the past year and a half I haven't followed their work super closely so I'm probably

missing some details but I think there's been a bunch of Innovations on like data format and like ways of using zup pass more easily as a developer so they're just continuing to like I think make it easier to use ZK kind of friendly data and um develop with it yeah I think I I had it I mean it was kind of used as I think you actually needed to get it for your initial ticket to like get in or something I mean I think there's maybe ways around it but it was pretty integrated into the experience I also know that I like I lost it at some point um and then for Dev Dev connect you needed it again there was a way to do that so I was able to get it back which I which I yeah I don't actually know how worked under the hood but it's actually

something that we would love to talk about more but with a lot of this stuff that you have like true ownership over it's actually like a very fragile system because you need to remember in the case of ZK data it's less of like a signature private key but more of like an encryption private key uh which is usually derived from a master password and if you forget that you're kind of just screwed um like you just like cannot recover your data and so yeah it's pretty interesting like these systems can be a lot more fragile cuz there isn't really easy data recovery so I guess is this a good time to ask what is cursive yeah yeah we are a team that is exploring new ways for humans to discover and deepen their connections so we want to make experiences that align

with user empowerment we believe that users should have control of their experience and we think programmable cryptography is a really cool toolkit for enabling this giving users ownership of their data giving a lot more Choice over what types of front ends and applications that they might actually be able to use enabling new ways for them to express their identity and ultimately this is all in the aim of connecting people and bringing them together is cursive like a lab is it a company do you have a product I feel like this is something you've actually probably heard a couple people ask I think I might have asked you this a couple times too but yeah what is cursive exactly I think we're currently transitioning from closer to an R&D lab to more of just

like a product team um I think like we have a lot of good lessons on the research front and like what is kind of possible and you know more difficult with cryptography so I think 2025 will definitely be a very like product driven year for us well we were in a very interesting position to begin with because I think we had a lot of product oriented thinking in our team but at the same time a lot of the technologies that we wanted to explore and use simply were not ready or were not packaged into user facing form and so we had to spend a lot of time not only iterating on the product side but also just fundamentally seeing what is possible what can be done

with these tools so in a way we were forced to do both um but I think we took a lot of really good learnings from that nice I wanted to ask what's your your approach to designing these products right I've seen a lot of teams who fall in love with the technology and say like this is great I want to use it and I get a sense that you guys do it the other way around right I mean I think there's a there's a healthy amount of that as well like I think Andrew and I definitely get nerd sniped and like just want to figure out new ways of using stuff we think a lot about constraints in our process and a lot of this is spearheaded by like our head of design Rachel like just creativity really I think is easier to access when you give yourself a box to work within

instead of being like you know the guy the limit I think one of those constraints is like around like human connection and data ownership like what are sort of the existing problems with with systems that we have right now um and in what ways does cryptography help in what ways does it make the ux bad like I think we're also very honest as a team and very pragmatic I think we understand that like it would be lovely to have the purest system with like you know like trustless or something close to that I think often that just needs to be balanced with like what users about and like how good the experience can be so yeah I think that's been the kind of the process in the past I think moving forward we're looking to be a lot more user Centric just like focusing on a specific Persona like kind of more

traditional startup stuff um given all the knowledge we've acquired on different cryptographic Tech and progress that way but that's a little bit more new um and I think is going to be like a 2025 Focus yeah I think in the early stages of any new technology if you want to be both on The Cutting Edge and also things for consumers there is sort of this dialectic between product and Technology right you you build some things you give it to users you see how they feel uh but at the same time you have to turn to the technology and say oh can we build this and also what else can we build and this new thing like do people want that uh so I think it's difficult in the early stages to be too much in one or the other uh but slowly that'll change let's do a rundown of the

experiments that you did um in the last year and a half or so um and and then yeah because I'm really curious to see kind of like what you learned maybe with each one did you do something after zalu once cursive was founded before Dev connect or was Dev connect sort of the first time that you guys went out as cursive actually I think at devc connect we weren't even properly called cursive but it was the same oh you weren't okay team I think cursive itself was launched at Denver okay um but there there was some experiments in between there we'll just quickly run through 2023 stuff yeah um and then pass off to Andro for 2024 so at Zulu basically we did a really simple experiment with NFC cards where

there wasn't even any ZK involved but you could basically collect like kind of like a private po app where an NFC card had its own private key and it would produce like unique signatures upon every tap and so you could basically assign an NFC card to a specific event at susulu like let's say attending a town hall and people who tapped that would essentially get this unique signature it would be stored privately in their zoo pass and then theoretically they could use it later you know to prove the attended either just with the signature itself or using a a ZK rapper we didn't actually get around to doing the ZK step it was mostly just like kind of maybe more on the side of like verifiable credentials than like um zero knowledge proofs but yeah yeah I think

we just wanted to create more data for people to make ckps with that was kind of the problem we identified during different workshops with Rachel and Alia of like yeah this Tech is cool but like no one can use it because like there was like ZK email and TLS notary vaguely existed but it was too hard to use those so you couldn't make proofs about anything I guess it's also on your side lay the groundworks for what's to come next right like all the infrastructure that you guys are going to reuse absolutely yeah like once people have a you know a little supply of signatures you can start to you know do a lot more fun stuff with that yeah so that was that and I think like the next step was basically adding these signatures and so this came around yeah like late 2023

this also when Andrew joined the team um which was an amazing level up for us oh my God like we were just suddenly able to do so much more cuz Andrew is a of nature yeah basically we experimented with this concept called quests which is basically like a userfriendly rapper of CK proofs so essentially like we create a quest which was I think the best example was like oh to attend this event you need to meet like three out of five of the organizers so the idea here being like instead of there being like a fixed invite list it was just sort of proof of knowing the people involved cool and instead of calling it a DK proof we wanted to experiment with different terminology and wrap it better just so you didn't need to be a ZK person to like get it um and I think that was like

a a really fun experiment I think we did a lot of interesting things like event invites and scavenger hunts and and and we also iterated this uh at eth Denver just a much larger version of this in collaboration with iyk like another NFC team main takeaway from that though is that ZK felt like a little bit of a limiting toolkit in order to build like full user facing applications around in that like a lot of what you can do is roughly like privacy preserving Access Control you can basically like prove you meet some threshold and then get access to something and it turns out like this often kind of just looks like tickets or like gating some like online content and I think we felt a little bit limited by having that be our only kind of cryptographic tool which led to like a

lot of future facing experiments at like zk1 and onwards cool at Dev connect it was that where you were actually doing those like you must know three out of five of the organizer type experiments yep uh we ran a few of them for different like parties and events um cool I I was not aware and sadly didn't go to any of these I did not know three out of five of these organizers clearly I guess so I guess so yeah we had like um 200 cards produced it was actually quite a fun Sprint like we basically were able to add um Baby J curves into these um cards like so basically they're Java cards so you code them with like kind of Jank Java like not full Java like partial Java and we were able to

get some valid signatures going and so we were able to build a full ZK experience around it and yeah there only 200 of them so I think it was like kind of hard to to get it to everybody H it's okay I I forgive it was a big event no worries um but it's sounds cool I'm kind of sad I I missed that part let's fast forward then to eth Denver 2024 because I mean I I really got to know what you guys were doing in at zk1 which was April so just like a month or two after eth Denver what had you actually done at eth Denver yeah I think two different things I would say the main experience that was available to every attendee was related to completing these ZK quests proving that you met certain people that you went to certain

events or uh had certain experiences and being able to claim prizes as a result of that but at the same time we were also working on new experiments specifically related to private set intersection and this was using a bfv library in collaboration with John maaya who's a longtime collaborator of ours and I think what we've noticed is from all these activations and all these experiments it takes a while to take a new technology in programmable cryptography and actually be able to productionize it successfully so at Denver we first played around with this uh bfv PSI

toolkit and Vivic spent uh multiple allnighters trying to get it to work um while we while we were building out the main M app experience as well um and I think we were able to make a proof of concept but there were still some discontinuities in the experience and fast forward to zk1 that's when we actually were able to deploy it successfully and it basically worked perfectly during the event and we' noticed like it really did yeah that the the lessons we learned from trying to get it to work and it uh struggling to work those carried over um and eventually we did actually like do it

successfully but at the same time at zk1 we wanted to also experiment with with folding specifically regards to being able to prove you met multiple people um each one of those being an an individual proof uh but being able to fold that together into some sort of Spotify WRA like experience for your whole zk1 journey uh these were all the people I met these were all the speakers I met and kind of generate one big folding proof related to that um and we didn't learn our lesson at all uh deploying Technology For the First Time still is a huge trouble uh specifically with regards to folding we ran into issues with key sizes basically having to download very large keys and

you can imagine like trying to get I don't know what was it like three 300 attendees 400 maybe more there's 5501 yeah 550 attendees to download like tens of megabytes of like keys on on conferen everyone on the same Wi-Fi yeah I don't know if everyone did it in ck11 I feel like it was the first time we worked together and we did announce it at the beginning of the event but a lot of the community weren't used to it by zk2 I think the percentage of attendees using it was actually up yeah but yeah but I I do want to touch more on the PSI or private set intersection experience uh because I think this was something that we learned that was actually quite successful um so

specifically what this allowed you to do is you have a list of all the people that you've met at the conference these are represented by their signatures signatures from their public Keys rather uh you have a list of interests or things that you're interested in and when you meet someone you can basically privately figure out what you have in common um and at zk1 we found that a lot of people were really interested in trying this out it got received pretty well at the event and people were like oh my God like I can actually meet someone new connect with them instantly and all my data stays on my device and it's all private so I think that was a really good learning from that event that even something pretty small like PSI it seems like a very simple primitive uh can be quite expressive and contrast that with

heavier things like full-on fhe like let's say you want to do fhml or something um this is quite complicated to run this takes a lot of computation and a lot of memory and it's quite impractical for a lot of day-to-day consumer experiences in contrast something like PSI is already practical today it works it works on your mobile browser and it can actually create new social dynamics so I think that was a very important learning for us I think like one other metap point here is like this is our first time trying like multi-party stuff where you took data from like multiple people and did some computation um ZK is inherently a very single player Endeavor there's like a verifier obviously so there it's kind of a multi-party computation I think some

people would get angry at that but um like the person with private data is only there's only one of them um the ZK 11 experience opened a lot of new doors in terms of like okay cool like what are the new social dynamics we can explore when multiple people have private data involved and it's a much less explored design space uh than ZK so that's one meta point and just building off the other thing that Andrew said like yeah I think like a repeated lesson or like kind of a repeated pattern for us is like we very ambitiously try to do some new technology at event like X we like completely fail it like blows up um like and did mention like for for Denver like we tried putting in like the private

seter section it literally like broke the entire app mind you we have like 12,000 NFC cards that we giving out like in 5 hours oh my God it broke it yeah the night before it 2 a.m. it was 2: a.m. we were doing some last minute tests before we went to bed and then it just totally broke everything and so we had to like revert the commit clean up a bunch of stuff test another few hours the moment on your face when you realize yeah there's no sleep for me tonight oh oh I love the term programmable cryptography cuz it's it makes you forget how hard it is to program this stuff oh yeah that's yeah honestly it's true it makes it seem a little too easy um but I think now we have a good awareness of like push to try to get it for an event

but like don't be too sad if it doesn't work out it'll probably work out for the next one and just a word of advice that like the stuff always takes longer than you think it will especially if you're the first person like actually getting into the weeds of this stuff and trying to put it on like a phone or a consumer device yeah it's great to see like your cryptographic like pallette sort of expanding but at the same time you're helping everyone else right CU now that you guys have tried this someone else can come in and try like more PSI stuff yeah totally that's definitely I think one of my favorite parts of the field is how open source friendly it is like how much of a collaboration it all feels like we're all just kind of like one big happy family like you know at least some

Corners some siblings biger but it's a big family it's a big family like everyone's building off each other um and I really like that like it's it comes in contrast to like ml for example which is like a lot of Clos Source work happening and then you see like deep seek come out and everyone's like this is such a gift to to manine it's like well yeah we've just been doing that in in cryptography land it's yeah I think in programmable cryptography a lot of people understand that siblings might fight but we're all fighting a much bigger battle the grand scheme of things like people are either gonna adopt this stuff or they're not and it's gonna take a lot of effort to make it happen there's a lot of inertia and so I think it really is important for teams to collaborate and understand that either everyone wins or no one

wins there is one thing I wanted to bring back about the way you did PSI so PSI has sort of a history and literature of being a pure like two-party thing like you have two parties finding an intersection I think in your case you guys did something quite interesting where you had like a server helping users compute their PSA could you tell us a bit more about how that worked yeah for sure essentially what we did was probably a little bit too overpowered of a technique but it like works pretty nicely for our use case which is uh multi-party FHA or like threshold FHA so what goes on here is like uh in this case it's two parties so let's say like me and inro are doing a private set intersection we essentially like together perform an initial MPC where we

create like a shared public key so between us we can see the public key but we neither of us knows the corresponding private key it's kind of like a split between us and so with this public key what we can do is we can encrypt our data fully homomorphically to the shared public key so now me and Andrew have like encrypted versions of our two sets of contacts interests whatever and so now what we can do is we can run a comparison in F and like usually this is actually like a very you know costly operation so the nice thing here is you can like offload it to a server um in a way that doesn't actually leak any privacy the server essentially just doing a bunch of like ffts for you and then gives you back the encrypted result and then at the end me

and Andrew come together and we do a uh decryption MPC where essentially we decrypt the result of this computation um without ever actually surfacing the kind of shared private key we just just decrypt it and thus only get the result of the computation um and this technique works for Stuff past private set intersections you can like do any computation with this method but you relieve sort the user devices from the heavy workload exactly that's what I found super interesting yeah like essentially the user is just responsible for encrypting and creating this shared public key uh which tend to be very light operations in general and so yeah you basically do the setup and then a server just handles all the heavy lifting and also when we

started exploring PSI we didn't have a good sense of whether or not PSI was going to be the main thing that provided a lot of application phasing value so I believe that a lot of the tailor made PSI schemes are specifically useful for private set intersection but using multiparty FHA gave us the flexibility to say oh maybe we can play around with different types of Primitives here as well maybe we can expand this to different things uh if they prove to be valuable yeah I think another shout out here to John maaya again um like without him like none of this would be possible specifically he

has a really great toolkit called Phantom Zone it's like an MIT licensed um fhvm uh different licensing to other sort of F tooling and it basically is focused on this multiparty FHA techniques so would highly recommend any devs to check that out um really cool Library can definitely build some fun stuff with it I remember when you first approached me for zk1 I want to say like I was really excited about the ZK the fact that there'd be kind of like a tool on site but I was also a little bit wary because in my like in the past in the past long before I met you um I had sort of seen a lot of these experiments go very very wrong in events and when you're running an event obviously you're under a lot of pressure in general you

want to make sure that anything that you're bringing will add to the experience of the participants so I already knew you Vivic and I guess that maybe is why I took that chance but I remember after the event being so happy that we had but it really was unlike you know I really didn't know what to expect going into it and so obviously when when we started speaking again about zk2 I was like very very excited to work together again um I know that between ZK 11 and ZK 12 there was another experiment that you ran though so let's talk a little bit about that Frontiers yeah I I'll I'll step in real quick and not assuage any of your fears um maybe you didn't know this but um so quick shout out we also worked with uh Jack andan um from Mach 34 they work on a

bunch of ZK programable cryptography uh tools as well they helped us integrate folding into the zk1 experience and at the end of the day like folding worked end it just was like very poorly performing due to key sizes during the actual event um we got it working in production 20 minutes before the event started that totally does terrify me you never told me that I probably I probably should not have told told you that but now you know amazing come down to the wire wow one last dk1 anecdote for me is how 1 hour before the end Anna realizes that not only is there like all this cool ZK

and PSI functionality there's a leaderboard yeah that's true so only is like I cannot lose at my own event so she's running around tapping everyone's cards right and I think you won right I did win in ZK 11 I it wasn't just that I I cannot lose my own event I think it was just that I yeah I realized that no one was ranking no one was competing and so I started to get into it um for zk2 we did fix that by actually offering incentives and bringing back the concept of prizes which you had had at eth Denver and we didn't think of for zk1 but I know yeah let's talk about that that in between experiment with Frontiers and this hiring experiment because I thought of that as like when we started talking about ck12 you were telling me about this it was so I mean

it seems like such a cool use of this of exactly these tools that you're that you're working with so let's talk about that yeah there's actually a really nice tie in here with ck11 which is that um Frontiers is like a paradigm sponsored event it's uh like focused on open source rust and ethereum uh applications and georgeos is like the main curator of that event and oh nice he actually first saw our technology at ck11 um and that's why he reached out he reached out in our public Channel he was like let's do something um and then we were like sure not really expecting it to go anywhere but um yeah ended up being a really good collaboration and so also appreciate youo like digging a bit deeper into multiparty FHA cuz that comes into play actually a lot more in

this Activation so specifically we wanted to take that toolkit we were we were using for just private set intersection and we wanted to expand it to like a more general purpose function and this is right around the time where John maaya uh opened up Phantom Zone as like a full I guess not a full VM it's closer to a circuit Builder where you could just write more general purpose uh multiparty FHA uh code and so we basically were looking into matching as like a core verb or core action of like how can we match people together based on private information in a way that like is more efficient or gets to use kind of like more data about people and in the context of like a professional event like Frontiers we kind of stumbled into hiring as an interesting example I

think there's two main interesting private information like pieces here the first is like salary so um as a recruiter there's some like usually like upper bound of salary or like ideally want to pay somebody and as an employee there's some lower bound that you're like willing to work for and you'd be excited to work for and so like discovering overlap there without having to give up that information we thought could be really interesting um so instead of there being like a lengthy negotiation and potentially even just figuring out later in the process that this isn't going to pay the yeah just it's like it's not enough basically you can sort of like frontload that without actually giving up the information and the second bit of privacy that I think is useful is like as a job candidate if you are already

employed somewhere but maybe like you know are trying to figure out what else is out there or maybe are unhappy with your job and like want to you know start applying other places you can't just post on LinkedIn like I hate my job like get me out of here and so we were thinking yeah like it could be interesting to basically like privately attest to the fact that you're looking for new jobs but you only get notified of jobs that kind of match your ideal circumstances so you can kind of like it's almost like manifesting in a way you're like kind of like manifesting your ideal job like the salary you want the people you want to work with and you're just putting it out there and only if there is kind of a match on those characteristics do you ever get notified and yeah we ship this end to end um using the same exact scheme where

the sort of like heavy lifting of this computation is done on a server the users are just responsible for encrypting their data and yeah it was it was like successful I think like I got a few like okay I put in a fake like application cuz you know I'm happy at cursive but I wanted to test out the feature so it's like the job market catfish I think exactly which actually raises an interesting point around verifiability that we can get into but for now I basically put myself as like the like ideal candidate like willing to work for like 10,000 a year like willing to work like 60 hours a week you know like just the most ridiculous ends of each of these Spectrum so I could ensure some matches yeah I got some job offers you know or like not offers but like some

things that matched and so yeah that was a really fun experiment I think exploring yeah what you could do with a bit more than just like this like you're really replacing the recruiter here too like usually you have a quotee unquote trusted third party who has all of this information that recruiter will know that you're looking but you're employer doesn't know they'll know often your salary bounds but like the potential hiring company won't really know exactly what your your your range is and this this is sort of like directly bringing the two parties together in that private way where they can find each other that's so cool yeah absolutely like I think this comes up a lot in our Explorations is like a lot of this stuff just tends to be much more direct with

the consumer and you can cut out the middleman and it was interesting like there's a tension here where we talked to some of the recruiters at the event about like what they thought about it and in some ways it's like well you're kind of doing our job you know it's like in some ways it's like kind of yeah not synergistic but I think one positive framing that one of the recruiters found is like actually it's nice because someone like georgeos can just put out a private job description and he himself can match with really good people instead of their needing to be kind of like a middleman and I think it can actually create an experience that's like a little bit less kind of extractive or kind of I don't know like clearly there's a bunch of like recruiter sharks coming around trying to hire people it's less of head hunting

and more just like peer-to-peer discovery of of good overlap and it can just be done directly between Engineers yeah there's it's interesting there's like positive and like negative framings for for Recruiters in this experience was there also like a communication feature where people could actually send something through this or was it just to the matching and then they had to kind of find each other elsewhere I think we just like once you match you just got their telegram or their email like you could just choose how you wanted to be contacted and then you could just reach out on those platforms okay but is this is this still like you're going up to somebody and like tapping a card are you connecting Are you standing in front of the person when you realize it's a match or is it like there's a pool of people and you're getting match kind of

willingness yeah in in this experiment you actually had to go up and and tap people and build your own social graph and I think the idea there was that if you met people in person you would be able to talk with them have sort of a deeper relationship and it would be more of a warm referral I think this relates to the question of data import which is a big question we spent a lot of time thinking about in our work because a lot of these matching experiments work well when there's really rich data available about the people who are getting matched and so in this case the data here is the social graph as well as the form inputs for your hiring preferences and you

could imagine that in an experiment outside of a conference you could get this social graph through other means be it maybe like ZK email from your calendar invites ZK DS from maybe your LinkedIn connections you could find other ways to import this data and you could also perhaps do that with uh with Job searches you could import data about where you worked in the past for example and have that be filled in automatically and so I think being able to easily import data is a question that's very relevant to making this experience richer for matching and it's something that we have to spend a lot of time getting the ux right because people filling out forms uh and

having to manually meet people in person to build that social graph that's not going to be the long-term solution and so still an open question and it kind of addresses the the catfishing concerns right like if you have verifiable sources of data yeah definitely I was curious was this whole experiment still built on PSI only or did you need other Primitives in there yeah so we actually basically just wrote a full Boolean circuit um specifically under the hood it's using few F fhw which kind of takes like Boolean circuits as input the the main kind of new functionality we needed was a comparator like a greater than because we wanted to compare the recruiter salary upper bound with like

the candidates lower bound and everything else ended up just being like bunch of ANS and oras and stuff like you know is there overlap on a experience level like I only want to hire people with like a bachelor's or high and just making sure that that like take the end of that bit vector and the candidates bit Vector is there like overlap basically cool so I just remember when we were planning when we were talking about the zk2 activation we had actually thought a bit about this hiring experiment is maybe also something we could run there um we decided against it I remember us kind of going over a few iterations of what like what could we get out of this moment with these people what can they get out of that as well like where where would

they get a kick out of this I think what we ended up landing on was having sort of preferences again because in ZK 11 I just remember it being like number of talks and number of contacts at ZK 12 we actually had something more like what are you into what do you like what do you not like or or I think the question was which of these are overrated but people got confused on the scale was like was three overrated or was one overrated so everyone got a little bit confused how to vote for it which led to some great conversations I I thought um that's on me but yeah but maybe share a little bit about how this differed what what were you using for this I think zk2 for us was more of an

ideological progression I think dur in the actual app experience we did do more of these hot takes finding more Rich ways for people to actually connect with each other based on more meaningful data I think what we realize is the questions you ask people and what they're responding to really matters you want two people to be able to meet each other and when they learn each other's preferences actually be like oh that's cool this is a conversation now or I'm really excited uh why do you think this so I think we we learn a lot in that direction um but more specifically we kind of had two new Concepts that we were playing around with which were just framings of things we had already done

specifically digital phermones and narrow casting um and narrow casting also being relevant um to Trinity which was a new MP scheme MPC scheme that we were looking into at the time maybe I can talk a little bit about digital phermones and and Vivic you can take uh narrow casting with Trinity for us we were thinking about okay like what is actually going on here we have this ability to Match multiple people on private information sort of build more meaningful connection through these interactions how can we describe what is happening here and I think what we ended up on is this idea called digital phermones specifically phermones being these

discret chemical signals that you're constantly sending out in search of human compatibility or a mate or whatever it might be and digital phermones are the digital equivalent so they can be applied to many different settings it could be online dating could be private hiring but crucially what ties this all together is the idea that they are private and safe so when when I have like my my physical phermones like this is not something something that I'm just yelling out uh or Screaming into public space these are much more discreet and these are things where okay when you

actually interact with me then they get surfaced in passive ways and I think digital Fones is kind of a similar concept where these are all private signals um they're all based on potentially sensitive uh data that remains private to you but the only time you actually realize something is going on is if you are connected with someone or if you do meet someone that matches your signals and so I think we we thought this was really cool because now there's this way for people to just be walking around going going about their daily life potentially walking around a conference or event just doing what they are normally planning on doing but behind the scenes all this matching is

going on and all this compatibility and connection is being found but none of it is explicit none of it is being done on a big server that sees everything for example yeah this reminds me of um contact tracing yeah when sort of the pandemic broke out you know there was sort of Two Worlds there was dystopian world where one government would put out an app that tracks everyone and knows all the interactions of every people and there was sort of the utopian version of that a lot of academics were pushing for and it's something like this where you would disperse your scent if you were positive you could send out a signal and those who picked up your scent would be alerted but privately like only they would know so I'm curious to see where this goes like what other sort of use cases we can find for this yeah that I think probably was the first

kind of like widescale deployment of something like this and then the hiring thing we just talked about is very much kind of like also a Proto version of the pheromone here literally just being like I want this job you know yeah like I like thinking of it as basically just like a private manifestation kind kind of it's like you kind of put out what you want for yourself or like who who you are and where you want to be and you can sort of just discover uh matches and overlap you know if this is to be something that is like fully akin to like natural pheromones like chemical pheromones I do think cryptography is the right substrate for it in that it is like very neutral it's very peer-to-peer there's no dependence on a centralized server mhm like I think this is like one of the rare things where a te does not of fully subsume it because it I think

does fundamentally change the nature of something like this where it's you do have to go through some service that is like providing this matching versus it being like yeah literally my phone is letting out signals to the people around me or like on an online platform and they can just parse it so yeah I think this was a really fun kind of Framing and wrapping for a lot of our technology that made it easier to explain and think about and we actually picked it up at dweb uh decentralized Web Camp uh in California just from some conversations there uh with kind of more like peer-to-peer local First Technology people oh uh uh another concept that you've been very excited about is this idea of local first Hardware so I think this is a great fit for digital fairmon

because these are kind of things that just need to be happening in the background and you could run it on your phone but the problem is Apple is very restrictive about the things you can do on a phone um that is just their philosophy I think everyone who has developed on iPhone knows this and they are one of the most fun products to use as a user and one of the worst companies to deal with as a developer as many people know yep exactly um so there there's limits to what you can do um with an iPhone with existing devices but I think a really cool concept is having open Hardware that like somehow is connected to your sensitive data but keeps it securely on that hardware and

you can actually walk around maybe wearing a pendant or something that's just constantly sending out these signals to other people um and finding compatibility that way just in your normal day-to-day life and so I think that's a very like interesting area of exploration if anyone is very knowledgeable about hardware and wants to work on this stuff please let us know cool so I think another concept that we introduced at zk2 uh was narrow casting which essentially is like the opposite of broadcasting you can basically pick like specific criteria that you want your message to be received by and kind of the cryptographic version of this is that like literally It could only be decrypted by people that fit this criteria so like witness encryption is

kind of like the purest version of this but you can build something with similar affordances using a mix of like ZK and NPC another related result to this that we published for the first time at zk2 was a new scheme called Trinity which was developed in house um basically mixing like garbled circuits uh plunk and then a very recent result using like kg witness encryption really beautiful paper would recommend people to check it out and yeah like one really nice affordance is that you can build something like narrow casting really easily because it has really low rounds of interaction between in this case the two parties this like a general flaw of MPC if you want to build apps with it

there's like usually a lot of back and forth and it's pretty annoying for an app developer to deal with if you're trying to make a consumer app just for the reason that people put their phones away like literally just like once the phone goes away it just like no longer can respond to requests as easily like apple exposes some things was this part of the motivation for your server aided NPC model yes absolutely like that's another place where multiparty F shines is you can delegate a bunch of stuff to the server and so you don't need to be online for those phases um but ultimately it still has like about four rounds of back and forth interaction cuz now you need to send something to the server server needs to get back to you you need to then do this decryption um whereas Trinity is basically just one round of interaction

like I basically just send like an encrypted email to somebody and they can decrypt it if they match some criteria and the other nice thing that Trinity adds in is is actually verifiability of input so this relates to Nico what you were talking about with catfishing like this is kind of a general problem with MPC is like you know if the inputs aren't like verified or signed in some way then you can basically say whatever you want right you can sort of like discover the other person's private information by just like repeatedly querying them and so Trinity also basically adds a really like clear tie between like zero knowledge proofs which can prove various like properties of your input data and like a multi-party computation which in this case is done with garble circuits cool moving on to your next events Andrew you and I jumped

on a call after zk2 and you were at Edge City uh were you running something similar to like zoo pass in Edge City what was what were you working on there I think first off shout out to Janine and tore and the rest of the edge City team for letting us do something throughout the whole month at their event I think they were very gracious and kind and uh supporting all of the hitches along the way as we were developing this um at Edge City Lana we built an app called cursive connections and the idea was to better connect people within this specific Spear of edge City Lana um and so some things were similar related to people being able to form their own social graphs

from NFC tapping where everyone had a wristband um but some things were also different in that we had matching that was more specific to the edge City community so one big thing at Edge city was the tensions game uh I think Timber uh ran that and the tensions game was hot takes where you could fill out your opinion on these hot takes and when you actually met someone you would figure out whether or not you agreed on some of them but more importantly which ones you disagreed on and within the edge City Community this was a big topic of conversation and another interesting experiment we tried out at Edge city was we wanted to see if we could match

people within a specific event so we chose the Halloween party that they were throwing and we wanted people to be able to walk into a room um kind of upload their their preferences here and within the event basically find who else who was there that they might be compatible with or might want to spend more time with after the party like a dating app was that a dating match we framed it more we framed it more as a friendship finding okay but I think this could but really it was dating dating in certain context okay yeah yeah we we're just surfacing the the connection however far that

takes you cool um and I think there there are a few interesting takeaways here um number one is just a very practical lesson of people don't want to take out their phones at events oh yeah and I think this is actually what inspired some of the hardware directions as well uh the idea that you can just have Hardware that's just like passively there that maybe lights up when you find connection uh that would make much more sense at this kind of event as opposed to people needing to check out their phones um and another related learning uh which I guess we'll we'll talk about more is specifically we wanted a lot of the computation to be uh non-interactive so we wanted to be able to surface the matches directly within the event without people having to check their

phones and so this naturally surfaced uh the idea of using te's because the problem with multiparty FHA is you do have this interactive decryption step so we thought of this example as one where you could enter room and you understand that the data you share within that room or within that ephemeral context will go to secure Hardware because it needs to match you non- interactively I was curious actually um you're just just mentioned tees so far most of the experiments have been purely cryptographic purely like software mathematic we have these hardware tools some trust them some don't they kind of fulfill a lot of the functionality we'd want to use cryp like cryptography for and so my question is as a design team

where do you draw the line between staying in pure cryptography or reaching out for the tees yeah it's actually a question that we've grappled with a lot this past year and there's a lot of different angles to look at it from one angle is like funding so uh we so far have been funded by privacy and scaling Explorations at the ethereum foundation with grants and so like they support programable cryptography research so a lot of our Focus has been in that direction because that's the community and ecosystem we've been a part of and it's been very synergistic in that way in that we're able to take technology that's developed in house and then find you know like product facing uses of it and I think yeah as as long as we're funded that way will always be a

priority to explore this stuff another angle here is like user first and I think like something that we have yet to identify internally and as far as I can tell also externally is like when people care about using the full cryptographic version of Something versus something like secure Hardware ultimately it doesn't really matter what like purist cryptographers think is correct if you're trying to build stuff for people it just matters what they think um at least if you want people to use your stuff if you don't care then I guess you know make your perfect Utopia I guess and I think like ultimately there's trade-offs to using this stuff and I think I think there's like a common phrase that comes up in cryptography Circles of like oh it's just dependent

on math and like for some reason in this world that's cool and makes sense but I think for a lot of people this is just intuition like that seems a lot less strong than just like here is a really powerful box that like can't see your things like I feel like that just clicks a bit better like at least I could see it clicking better and so yeah like one thing we definitely want to clarify more in 2025 both for ourselves and for the external Community is like yeah when do people care so running some experiments with tees running some stuff with pure NPC and just getting some user data on like how do you respond to the copy around this stuff cuz ultimately you know an average person is not going to verify a te attestation or like look at the verification key for this proof

they're completely trusting you and the sort of copy an explanation that you give them and like the sort of feeling the appics so basically like my answer is like to be determined what balance we're going to use for this and we want it to just be very user driven versus like what we think is right and yeah excited to share with the community the findings we get there yeah super curious to see results yeah like I think a lot of people are either committing to one because they believe in it but I think I've seen less of a kind of like let's try them both and see what happens yeah one way I've been thinking about this is in terms of the hierarchy of trust when it comes to data privacy and ownership so you can think about the lowest level of trust a user has in an application is

one where the application just like makes no claims to keep your data private uh and just like blatantly will sell your data and everyone knows about it the level that comes above that is one where the application is just like trust me bro uh I know what I'm doing I'm not going to like sell your data or anything um after that comes regulatory trust uh an example here in a different topic is like licenses like people trust licenses because you can like eventually bring someone to a court of law if they violate the license um and at the end of the day you like have trust in that system because of the regulatory mechanism that underlays it and then one above that I would put secure Hardware um where you maybe don't have to trust

the application developer as much if they do remote at to station if they like open source their code and everything but now you're trusting Intel you're trusting like State actors to not break into this uh maybe you're trusting an $150,000 microscope or whatever the cost is for that but then the the top of that or rather the second from the highest is cryptography um this is like trust in math and the real beauty of cryptography in the trust hierarchy is that you can verify it from first principles like Intel is is closed Source like or sgx CDX like you know you can't go into every single layer of depth and check it yourself but theoretically uh you could do the same with MPC with fhe and actually go in and make sure everything

checks out and that your data is safe um first principles plus a few degrees but yes yes yes um yeah more practical framing is that you trust a bunch of smart people with a lot of degrees to actually go in there and and verify this and tell you like hey it's good which is also why cryptography really just has to be paired with open source if you can't go in and verify it like it kind of makes no sense to begin with um but there is like a sneaky like I don't know it it technically belongs in the top of the hierarchy but it's like somewhat restricted to uh certain a certain design space which is just local first you can think about this in terms of llms processing your data if an llm

wants to process your data and it's running entirely on your machine like that that is arguably like the highest level of trust you know you can just turn your computer offline and you can just run everything locally and and that's great um but that's restricted to singleplayer applications like once you want to get into multiparty computation then you have to do either cryptography or secure Hardware um so I do think like the big open question from a user first perspective is what applications are best fit for which level of trust uh and how do users actually perceive these things and I think to answer your question cryptography is really important for uh very sensitive data anything that might involve like Medical Data anything that people really might not want to share um with a server I think that's where

cryptography can really shine I want to talk about the last event of the year I think for you I mean if we just go back this is I mean 2024 is sounds like it was just a marathon of events and activations for the two of you for the group um but let's talk a little bit about Devcon and what you did there there's like an art gallery I remember you telling me about this before you did it how did it go yeah main new technical thing we did was uh collaboration with yet another guest on this show um the folks from t uh to essentially do a kosnar experiment it's actually kind of a iteration of the ZK wrapped kind of in a way where essentially we wanted people to be able to like prove to us that they'd hit some number of taps or

connections at the event again using NFC cards and this sort of thing to produce signatures and we had a little prize store again it was like NFC Rings we had like NFC bucket hats we had bracelets we had the whole like the whole lineup NFC bucket hats I've never heard that NFC bucket hats yeah yeah I think I have one somewhere got one of the last ones it's just an NFC like a cloth an NFC tag behind um a label it's actually really nice and also setting up that clothing line was super fun that was a collaboration with myself and uh Tessa who is a Contractor on our team anyway um we basically wanted people to make proofs for us that they hit some number of connections and we want to experiment with delegating that uh again actually in the same vein as like multiparty FHA

where you would love for you know less of the computation to happen on your own device and for more of it to be kind of done somewhere else and so yeah what people would do is they would take the signatures they collected they would secret share it into three parts and we did a fun thing where uh three different teams hosted the coar servers uh PSC hosted one of them the PSC infer team shout out to them uh cursive hosted one of them uh I am terrible at AWS I like messed up the deployment once or twice so I was definely the weakest link in the chain uh but we hosted one and then hosted one and yeah like together you know there's some sort of like you know theoretically we could all collude and see your signatures but there's kind of a

reputation level thing which I think is really interesting of like we're all privacy focused teams the chance that we do this is low or at least it would kind of hurt our reputations if we did and so yeah we basically did this whole delegated proving experiment and it actually like was like over subscribed like I think we had a few T tens of thousands of proofs overall nice and and there's one point where I think almost almost 20,000 across the whole event and basically again like this is my fault where we were supposed to like replicate the nodes so that there could be more servers making proofs and then not only did I not do this but also like two of our nodes broke so there's only one node on the cursive side and then eight for PSC and Tso and so we're completely

bottlenecked by cursive and then eventually we fixed it and it was fine it cleared out the queue of proofs that was waiting but yeah one of the coolest things about deploying things in the real world is that you get very direct feedback loops and the way we had set it up was that you would make these proofs to show how many people you've met and you could get some of this NFC merch by showing these proofs and so we realized that the infrastructure for generating coars was not working when there were tens if not a hundred of angry people standing in front of our booth asking us why their number wasn't going up oh no that was a fun day I love that

you you put yourselves in this sort of line of fire but it's just so it's I mean it's also you get like the visceral feeling I feel like after that you go back building and you really you take it in you're not ignoring that feedback I feel like it sticks with you no I think we're really motivated by that as a team like collecting it and also acting on it um I think I mean we do a lot of just like Ivory Tower stuff as well it's hard not to when you're trying to figure out what to do with this technology but we try to like be on the ground as much as we can and we really like every opportunity we get we try to like make the most of it reflect on it heavily and just like move forward I mean that's a huge number of events just there that we covered

testifying to what you're saying now and yeah again huge shout out to to folks um Ash Lucas uh the other Lucas um Daniel Roman a bunch of bunch of amazing folks helped both with the activation but also they just came to the booth and like helped volunteer and stuff so such lovely folks yeah I guess that ties into the the final part of our Devcon experience which was the cryptographic connections Museum that we launched in collaboration with a bunch of artists and coll aters and the idea behind the museum was just to Showcase all the work that has gone into cryptography help educate a lot of attendees who might not know much about cryptography and its history and also

tell people what's going on in the future and what to get excited about and we basically got a bunch of artists and cryptographers uh and everything in between to Showcase different pieces of work and I think the most exciting part of the whole experience for me was that Ash had set up a display about the history of MPC so she had all of these different fundamental papers I remember like uh the day before I had gone to a print shop in in Thailand and gotten like literally like 5 600 pages of papers printed of just math and they were so confused like what is going on here and the whole display was just these papers and and this like history

of NPC and a bunch of like golden frames um set up like an academics desk like a coffee mug pen paper all that stuff um and throughout the event like this was super attractive to a lot of the museum uh guests and two people uh Yen grth and Don Beaver showed up and they both had their own papers uh Gro 16 and and beaver triples um on the table and that was so exting seeing them walk over get to pick up their own paper and be like oh my God this is so cool that was definitely the highlight of Devcon for me amazing Nico actually was also a featured artist do you want to talk about the thing that together yeah yeah

did you guys printed out the ZK jargon gcod Yes actually it it was supposed to just be a museum piece and a prize but we actually had like three or four robberies that was actually like oh they all got stol okay wonderful yeah pretty much all of them I I think we saved one for you right I hope you have yours yes okay great it's right here Wonder wonderful the rest got got stolen so um if if you're out there listening to this we we will find you you weren't supposed to take them whoa wild okay so that wrapped up I guess 2024 for you when I spoke to you end of the year um you did sound were like I don't want to travel anymore although I think actually Andrew you

kept traveling yeah but we're now 2025 beginning of the year what's what's the plan are we going to see you doing more activations at events are are you actually planning on doing something else you you were talking about maybe a shift more towards product so yeah tell us what you're thinking yeah I think like specific technal directions are still being figured out but like definitely I think just in terms of like process of building I think trying to shift more to Marathon mode I think last year was definitely a lot of Sprints um of like something comes up it aligns enough with our directions and I think we were just really in deep need of like data like we just needed to like try things with people who weren't already pilled sometimes like at a conference like ZK Summit it's actually good to try

it with people who are pilled just to educate more than like necessarily get user data but something like Devcon leans much from the other direction and I think this year probably still will'll do some events and conferences just cuz they're convenient and we have good connections with them but I think building products for specific personas and just trying to like grow within the cities that we live in and like just try to get like further and further away from like ZK ethereum land cuz I think the sort of confounding factor in our space which is both a blessing and a curse is that like people are down to try technology for technology sake like they're like oh cool thing in the tech let me try it versus like this like does something good for me and most people operate entirely in the ladder Camp of like oh my God I don't want to have

another app on my phone I just only want it if it's going to like add to my life so it's a little bit like high level but just I think trying to get more feedback loops from people very outside of the space and like more targeted at like specific people like one example that comes to mind is like professional apps entirely so at people you know who are like LinkedIn power users or something like how can we make their kind of hiring job search experience even better and just trying to build a very like focused application around that uh is is like one example I don't know if we'll end up doing that but cool are you looking for the product like the singular product that you do end up building is that sort of what a lot of this was headed towards or do you

actually see yourself having a suite of different just more built out products I we'll see I'm excited to check back in like six months and and let you know but because we are in more of a like consumer and social area like I think what tends to be the right process in the space is like lots of shots on goal and you see what sticks and then go deeper with that so I have a sense that that's going to be at least like the short-term strategy is just building out lots of stuff giving it an Earnest experiment and then trying to have a conclusive yes or no and just continue and understanding that most of them will be Nos and that is just kind of how it is cool I can't wait to see what comes out of this I'm so exced excited for for cursive actually I really liked the

approach like from the start and uh yeah I want to see what next year brings yeah totally and on the flip side uh we're also always looking to collaborate with other teams on MPC FHA research uh any sort of technical research that might align with these directions we've had many successful collaborations in the past with different teams be it psse gaus ZK Emil uh a bunch of people doing really good work in the space so if you are doing any technical Explorations and you want to maybe see it integrated into applications user facing products please do reach out to us are you worried at all about like some of the patent stuff in the FHA world like does that stress you out because you're touching FHA a

lot and I'm just wondering like are any of the libraries you're actually using potentially at risk because of the more extreme patenting and less open- Source nature of at least the way fhe seems to be developing I think for the time being actually like multi-party fhe is no longer a key part of our tool kit just like right now there's a little bit of there's a few bottlenecks like in particular like public key sizes are very large it's a little bit slow it's like a little bit hard to use and I I don't foresee that changing in the near future or at least we have a lot of other options that we have to explore more fully before we come back to it so it isn't directly affecting us we were talking about how lovely it is that we're a big family and this feels like a

certain like you know Waring faction just like decided to leave and and make things harder yeah um and I hope they turn that around like you know as Andrew said there's an understanding that we're not fighting each other we're fighting like the rapidly encroaching like centralized zero privacy based world and that's we need to be focused against um and not against each other so it's easy to say that in theory but I understand if you're trying to run a business it can be easier to be like zero sum I don't know yeah I I do think this is a question of long-term trajectory if you want developers to come to your ecosystem if you want people to build upon the crypto systems you are building I think open source is pretty imperative to that um now I I don't claim to know

the specifics about the licenses and the trade-offs there and what it means for business but I do think generally speaking the more open been the better at least in this early stage of development um because the question is not which FHA system is going to win it's whether FH even stands a chance in the first place yeah good call well thank you both for coming on the show sharing with us the Journey of cursive all of the activations and experiments and the direction you see yourselves going thanks for being here thank you thank you and also thanks for the the two spotlights at zk1 and 12 like that's been so great for us and also just you know to I think build awareness and education for this technology has been inspiring for us like it's it's really

cool to see what cryptographers think of this cool thanks Nico for joining this one thanks for having me on and I want to say a big thank you to the podcast team Rachel Henrik and Tanya and to our listeners thanks for listening [Music]