Proof of personhood panel | Devcon SEA

[Music] hello hello great so we just had a nice overview of the three protocols um and what I want to do in this uh small amount of time we have is just challenge each one I think there's something to learn uh you know from each approach here and see if we can come to some greater synthesis so I'll start with Remco so Remco I actually wrote a critique of proof of personhood it's called compressed to zero the silent strings of proof of personhood and what I did is I looked at a parallel protocol to worldcoin a different authentication mechanism but they had succeeded in uh verifying Unique Individuals uh through cognitive tests um and they had avoided the problem of account Trading which worldcoin has had to Grapple with because each account has money and then of course there's an incentive to um you know sell accounts early on and so it's something you're grappling with they had solved that with this novel mechanism called sublinear identity staking which I'm happy to talk to you about later and but they were still left with this information game theoretic problem around puppeteering which is that when you have an account and it has monetary value it has a Ubi stream coming towards it which which is one of the goals of worldcoin um you create incentives for other people to try to control those accounts and extract from them and so in the kind of py way to put it is when you try to differentiate humans from Bots and use Ubi incentives you actually create an unintended consequence of humans trying to control other humans as if they were Bots right and so the control problem becomes an information problem and it's an information game theoretic ones so I I want to know how you think about this challenge because there are two goals that are very explicit with worldcoin one is both Ubi and a kind of rails for Democratic governance and even maybe even uh AI alignment but if if ultimately you know pursuing both of those goals ends up in coalitions and oligopolies and basically politics right um and one person one vote break breaks down very quickly in Coalition it's like how do you deal with that or how are you thinking about that so um we're at 16 million users now 16 8 million um in the highest tier or verified unique human and and in the process of growing to this scale and being active in dozens if not nearly a 100 countries right now you kind of see everything that people can throw at you and while the vast majority of people is honest there are some very dishonest actors out there and you have the uh um Byzantine General's problem of you need to make your resist system resistance against malicious actors um I think this is not this this is a as as you mention like this is for every proof of personhood primitive out there there is this issue um as I said proof of personhood is not a capture so there is this idea that you can delegate an account this is not something I think we should prevent but it is something that you need to give the user the tools to properly manage which means that it needs to be explicit it needs to be consensual it needs to be revocable and so on so I think the emphasis is on developing those tools empowering the user to control their world ID in the way that they want to control it um blockchains as a whole are not particularly good at the moment at these kinds of things because a lot of decisions you make on chain are oneway streets but there are a number of things we are already doing here um like allowing you to recover your account by going back to an orb one thing we've rolled out that is very powerful is using the phone's front camera to do additional interactive face authentications to make sure that the wallet hasn't been stolen it's still in the possession of the original human that signed up and so on right maybe I Canad comment there directly so I made in this presentation my claim that I think this absolute identity there are actually not that many use cases for it and kind of just to give the extreme example why why I'm making this claim is let's say somehow North Korea does get hold of one of those orbs and and can um uh can yeah properly authenticate they're all there 25 million or I'm not exactly sure how many people they have probably no one knows but 25 million people so let's just say in worldcoin now you go from 8 million fully verified to now uh 32 but 25 of them are just are are are people from North Korea where it's not even clear whether they um whether they have control of their key or probably it's it's somehow centrally organized that someone ultimately has full control of those those keys now you have 32 million properly verified actual humans question is what application for what application would that Set uh be useful so my claim is in absolutely most applications and probably actually this gas subsidy that you're doing is one of the few where I think it is uh useful but uh but in absolutely most applications I would say you need more context than just uh Being Human or I mean if you want to go even more extreme we can think of this Matrix movie where they literally I mean in The Matrix movie where they literally farming humans or it kind of have those Farms of humans that basically just sit there and yes they might have an eyeball you might be able to scan that uh but but but that's really not you what you want right and even just to add on the now right as long as there's money at stake I mean we have synthetic biology it could also be just sort of synthetic you know eyeball forms to hack the system okay hold on there's like a couple different things going through each other here um synthetic eyeballs don't work because with Biometrics it's critical that you verified it is actually part of a real life human body um if you get possession of an orb and you abuse it while the sign up process between the user is absolutely Anonymous and privacy preserving the orbs themselves are not their behavior is public and um there are methods you can use to actually revoke the people that's the the fake accounts that might but but they are no no fake accounts okay that's that's the second part if they are not fake accounts um then the system works as intended in a way and you make a very good point that proof of personhood is just a primitive and you need a rich ecosystem of things around it the example I like um a very simple one is let's say you have an onchain Dow um you have you want to implement one person one vote if you just do this and allow all of the world ID verified users to vote on this and your like social Community is only 5,000 people you can easily get swamped you don't have a very strong governance system that way so you do need to incorporate this in more complex systems 100% agree can I add something yeah I kind of like I get all the points but I think like the like forgetting the orb or the eyes and stuff so I agree that overall identities and I don't believe in Universal identities because like identities are have to be built for the governance purposes so we should like come up with Frameworks that increase the coordination quality among the humans and then decide the depth of the identity how it should be built when we talk about the proof of personhood it has like two Dimensions right one is the proof of uniqueness and I agree for the real world and especially in multiple cases just uniqueness proof and badge does not guarantee the quality of decision- making or coordination and then we have to rely on passports or like citizenship or the age or like all the other stuff and the thing is that like if we like for example for voting refer to the passports the dependency on this document is so big that we can generate the uniqueness Based on data and we might not need any type of Biometrics additional to it right and I think like in a way that's the purpose of identity so I don't believe in Universal identities I believe of defining whether we like fighting AI or or or we like fighting the misinformation stuff the only systems that are dynamically evolving and improving are the ones that have this kind of fluid identity aspects and not just the universal ones like on the example of Twitter and the community noes right so just labeling is just creating this very low barriers in terms of defining the permissions what you can do or what you can't do as a human or not and like you said yeah there's a kind of no Korea scenario there's kind of a dystopian scenario where we might be like creating market right for the eyeballs and stuff and like you got two kidneys you got two eyes and then what happens next right so well what you raised is an interesting point about governance right because in some sense the the the the point Martin was making is highlighting an enforcement problem right and this gets back to the information problem yeah right and so even if you do succeed uh in authenticating Unique Individuals in North Korea how do you actually enforce that and what you're saying is Well you need richer governance mechanisms we need to express identity in different ways yes right as a complement and if you just have that one identity system ultimately enforcement would require surveillance and of course surveillance results in everyone becoming informationally the same so I want to shift gears Martin and poke a little holes at circles I think it's a great experiment but one of the things I noticed on your slides is you know you're you're talking about currency and money I think of them almost as Community currencies and like nest and networks of cooperation building in a nation states and like you know even the globe but you know not everybody in a community should receive the same amount you know Blood Sweat and Tears and contribution right that it shouldn't just sort of be like an hourly rate that's a kind of um you know another kind of Ubi that has its own uh pitfalls because it's not tied to incentives and um you mentioned a 7% interest rate these things seem arbitrary and does they don't seem to leverage incentives or people's voice choice and stake in a community so how do you think about that yeah I mean on on the should they receive the same amount um I I would say um if a bunch of people together comes together and decides to use something as money then just that coordination effort to extent creates value uh to whatever they decide U is money and I would argue that the fairest form of Distributing this uh this value that has been created by just this coordination effort uh would be to distribute it equally among those that uh join this coordination effort super important to understand that that that that doesn't mean that everyone needs to work for the same hourly rate or it doesn't need I that's absolutely not the case so it's just uh it's just that the coordination effort is in a way or the the benefit of that is equally um distributed sorry that was a second question that I just you know uh non-equal distributions within you know acknowledging that there's different levels of I believe 7% that that's 7% right yeah so so um so that is probably the only parameter in circles that is um uh yeah well that that that that has been said and that that's therefore debatable and yes you could say everyone every Community should be able to said that themselves uh that would make coordination much much much harder um because why well because then you would need to um or in in in circles it it is meant to be so simple that I I I I issue circuit and then I know I can kind of trust that other account I can kind of um I can exchange my money to their money one to one um because we are issuing under the exact same rules if everyone could Define their own rules it would be much harder to negotiate of how much should my money worth uh uh compared to you to yours so I would say for a stable equili or the goal of circuits is to have the stable equilibrium where um where yeah trust or or the exchange rates between circles are are just one to one the goal would be to say c one cir circle is one Circle and and it's kind of abstracted away the fact that under the hood we have all those different just like in the in the in the dollar system right now um technically speaking if you have a dollar at One bank and a dollar at another bank those are not the same those are not the same dollars um but the system is built in such a way that usually it feels like one to one because if everything goes well you can exchange them one to one but if it breaks you will realize oh the the the bank the dollar at the bank that just went bankrupt is different from the other but the US and Euro have floating exchange rates why can't you say again the US and Euro have floating exchange rates no no for sure for sure for sure so in circles yes you or I mean the expectation is they are within a connected cluster the exchange rate is is one to one but yeah there can be uh different clusters where within the cluster the exchange rate is one to one but between the Clusters um there are um a floating Exchange not just rely on supply and demand no well it will be I mean I mean the price between those clusters will be defined by supply and demand and what we are actually supposing um in our upcoming um upcoming uh um yeah relaunch I'd say or or restart or or um or reactivation it's a continuation of the system is that uh people people or groups can back their very own circles with an Reserve asset they want so let's say Bitcoin dollar uh whatever they want and then yeah there will be a a free floating exchange rate for for those circuits so defined by supply and demand um but as long as people trust each other then then they are essentially creating additional um yeah one to one demand between those currencies maybe we need to go back to the identity uh well no but it's it's it's uh you're expressing identity in a different way and so it's I just want to dig on that but I want to turn to lha now so how do you how do you see yourself um and raro uh in this constellation of identity protocols what what like differentiation do you see that you're bringing I think we tried to use the both Sid of the world such as like passports as an instance first but then extend it with like graph like and capture more depths of the interactions and uh what I would say is like how we try to solve the identity problem or enable the buildup is very use case Centric and like I said the governance Centric the fact that like we understand like this voting stuff has to be done that way it gave rise to solving the uniqueness problem in a certain way but this doesn't mean that that same uniqueness should be reused like universally across other use cases and so like going at A Primitive level of these kind of Registries and the kind of building blocks and you having more like this kind of flexible framework I think that's the way forward and as well as we've seen mostly right on the kind of regulatory pressure or like this kind of collusion risk I think the permissionless and empowering of the users going forward should be like this kind of core principle that no one should kind of violate or cross the line for so we only have three minutes left I want to quickly do one minute each on where you know AI is a part of identity is helping us leverage AI to communicate with each other also to govern AI so starting with you Remco um where do you see the most immediate use case with worldcoin in in the near future um in world or World ID uh interfacing with AI will you know open AI uh for example Grant free access if you have a world ID um what is the sort of immediate near term goals on that interaction I mean I can't speak for open ey but we do share some outm in common so we'll see what happens there um regarding AI um I think I think we need to embrace the fact that it's going to be some sort of like human AI hybrid operator uh I I I'm a big believer of like AI is just making hum better versions of themselves so this is why I think it's so important that you distinguish between a capture and a proof of personhood which a different Primitives um with different use cases and we might not even be that interested in the capture use case what is going to be a very important Pro problem is like distinguishing real world sensory data from faked data and authenticity um I don't think proof of person alone is a solution there but I do think it can be an important part of a functioning solution that allows us to authenticate content and information Lasha any immediate applications yeah in my world I don't have this like as humans and AIS and kind of like the separate camps of the camp I more see and question like okay what is the first thing like I own multiple AIS so the First Dimension is how do I control it so how do we build this my me my Ai and of identifying these relations and control and then zooming out like all the AIS of individuals or like the systems how do they coordinate and what are the mechanism for us I don't know like lower down the computational capacity or like just shut the certain things down and just like abstract them this is a kind of a next level of the governance problems that we should be figuring out and that's where kind of the identity aspects will be born uniquely Martin I know you thinking a lot about AI yeah um I I think it's certainly good to have a robust identity um uh framework in in place that also is is is relational because yeah for the reason in the very beginning so I like to see a version of something like like circles that's completely permissionless as a base layer built on on on this um trust graph but then uh yeah projects like um DK passport or or or World ID that can essentially provide additional um yeah attestations what I would call it um to those open um base layer identities can bring us a very robust system overall um where humans can uh coordinate if necessary maybe against AI yeah will each circle have its own AI agent that's a good question whether a AI agents will also use circuits or not I don't know yet they could all right let's give them a round of applause everybody thank you [Applause] [Music]