Harry Halpin & Jaya Klara Brekke: Nym – The Next Generation of Global Privacy #426
14 Jan 2022 · 1:09:28
Harry Halpin, CEO and founder of Nym, and Jaya Klara Brekke, the project's chief strategy officer, join Epicenter to discuss building mixnet infrastructure that protects internet metadata from mass surveillance. They explain how Nym's architecture shields packet-level metadata at both the network and application layers, how the protocol integrates with other applications, and the incentive design that keeps the mixnet running. The conversation also touches on Halpin's personal experience with surveillance and his views on privacy as a public good.
Speakers
Topics
Transcript
so welcome to epicenter the show which talks about the technologies projects and people driving decentralization and the blockchain revolution i'm brian from crane and i'm here with federica ernst today we're going to speak with harry halpen and jaya clara brecke who are the ceo and cso of nim and nim is like a really interesting privacy project that's trying to bring you know privacy to crypto and and beyond that and now before we go into that just briefly on our sponsor so first of all we have tally so tally is a new wallet for web3 and d5 that sees the wallet that's a public good so it's a bit like a community owned alternative to metamask it all has all the features
of metamask but it's a 100 open source and under gpl v3 license and it's you know it's fully user owned with all the projects flowing to the community so the launch is coming soon and they just have a new version that you can check out tally community edition before the daw launches and you know if you want to get involved you should try out tally join the discord community and yeah just go and check it out at tally.cash and then the second sponsor charles wall so you know dexes are great but the problem is they're vulnerable to things like meb you can get front run have failed transactions high gas costs so
household tackles this issue and offers a new kind of trading experience so it's built by agnosis federica here responsible partially and it's a so it's basically a metodex aggregator so it aggregates different dexes so if you basically put your order in and then it matches the orders automatically and then if there's no matching order it just basically puts it uh in a variety of underlying amms depending on whether it's the best price so yeah um give cows up a try you can also have benefits like no gas fees paid for failed transactions and optimized transaction management for multi-second dolls
so go over to cloudswap.exchange and you can get started so with that let's go to our episode yeah maybe just harry and jaya do you both want to introduce yourself a little bit and you know how did you end up in this working in this crypto space and working on nim so um i could start i'm uh the uh co-founder and ceo of nem and how we started is essentially uh you know i was involved in uh to a large extent uh two very disparate things uh for many years i did my phd in artificial intelligence
at the university of edinburgh and during my phd and my work afterwards i realized what small amount of data can be used to de-anonymize someone particularly metadata um i became very concerned with this problem i was uh personally kind of under surveillance due to a large amount of anarchists and climate change activism and i ganged up with a bunch of friends of mine including jaya we worked on uh european commission funded projects including one to build a early version of nem which is a mixnet which is a sort of uh kind of system which can prevent nation state level surveillance and then uh when i saw various people uh
particularly like i remember the block stream raised very clearly and other raises we said well look we can actually uh pivot uh this mix that technology into a startup so by dance through red uh then polychain through it at a16 zed through in this year and that's how we got where we are today tell us about the things that got you surveilled harry yeah i mean it's it's it's been a pretty wild uh wild ride i'm happy to talk about more of these things in detail but why don't we have jaya do a quick intro to herself yeah happily um so my name is jaya and i got involved uh at nim um joined as cso
a few months ago before then i was working more on the kind of content now more strategy side of things and as harry mentioned um yeah we worked together quite a few years ago and were kind of in and out of various social movements even longer time ago um and so my let's say first kind of encounter with the crypto space was as part of one of these european commission projects namely one called descent which was where uh i met a few people and read the bitcoin white paper um for the first time and that was in uh probably 2012 so pretty early on and it struck me immediately as a super interesting proposition because um you know i think both harry and myself have
been involved in lots of movements that have been looking at questions of decentralization and power and technology for quite a long time um and here was a kind of new iteration of that that also started to look at the economics of some of these questions um so that was super interesting to me and um i ended up writing a phd that was a political analysis of bitcoin and ethereum looking at some of these dynamics um and then did a bit of work with harry on next sleep another european commission project which was the precursor to nem um and then i followed nem's work uh for a while before joining the project and yeah i'm very happy to to be part of this very exciting project how did you get in trouble with the man
oh well that's pretty simple it's very hard not to get in trouble with the man um you know uh and how what what happened this is this is all i think public record right now uh is we were doing a lot of activism around what was considered a fringe issue at the time which was climate change now it's somewhat socially acceptable to discuss it in public in fact there's even bizarre netflix movies on it and we were trying to draw public attention and concern to climate change issues and what happened is i had an undercover police officer called mark kennedy which is now the news assigned to me uh and uh their job was essentially make my my life very difficult as as an
organizer which they were very successful at doing uh they you know the police even harassed my phd advisor came into my office at the university i mean it was pretty wild there's a lot of uh at mit and all sorts of other places um and yeah i'm still on a blacklist so i have trouble flying um that being said uh you know the the the cop later revealed himself and the whole operation fell apart but uh that made me realize uh how damaging i think surveillance and can be in privacy violation so it was not only myself i mean to be honest i i got off quite light uh being on a back blacklist is not really the end of the world particularly if you run your own company
um but uh a lot of my friends really had their lives utterly destroyed um [Music] for years if not longer and i i think an important thing to realize uh is that no matter what you're doing so you may think that you're just sending bitcoin around you may think i'm just enjoying some free yields on a d5 project uh you may think what you're doing is uh completely innocent legal lobbying uh you know if you really anger any of the kind of forces that be be those corporate uh the undercover cop was paid by actually oil corporations for a bit or a government
um you know at some point they're going to want to shut you down uh and when they want to shut you down they will there is no law involved they will do whatever it is they can uh to surveil your communication and to uh destroy your uh relationships and to get you fired from your job and not maybe if you're in the middle east uh several friends of mine were put in jail and tortured and so it's uh it killed uh and so you know when you when you see the amount of devastation uh that surveillance has caused it seemed natural that we should have protect i mean also to be honest i worked for tim berners-lee and we saw that all the policy-based approaches the ending masterminds did not exactly succeed it's still very much in fact maybe even it was legalized
after the snowden leaks so so you know it was an obvious move to sort of say well okay are there any technological uh fixes and there's no simple technological fix is there any technological let's say self-defenses that someone can use against surveillance and it was very clear that when you're up against nation state level adversaries as i was um current defenses aren't strong enough and therefore i sat down with my old friend this is you know 2012. i sat down with my old friend george denises uh who went on currently stolen them has had a sort of checkered past with the company uh who was working on mixnet which is a technology that predates tor and he was
work actually with the tour founders on an earlier mix that called mix minion which was probably the mix that used by satoshi nakamoto even to hide his identity when he launched the bitcoin white paper that's because george i said look we should just revisit this and build a new mixnet because i personally want to see this technology out there and luckily george anya the rest of the nem team designed built uh first under non-profit funding and now under uh startup funding a new mixnet which we think is uh kind of what we would call best of breed against very powerful global adversaries who can watch every bite coming in and out of a network and it's not perfect you know it's not going to stop a lot of things that hurt me good old-fashioned uh undercovers who you know whatever
come to your house i mean obviously you're mix that can't stop that but i do think it it will be a a valuable tool in the future not just for activists but for everyone because again the main thing is when you get targeted you don't think what you're doing is illegal i definitely didn't think i would be targeted but nonetheless it will happen to you and that's why it's important to have these technologies built in by default and widely available to everyone like i think there is something so something really important to understand also about the difference between kind of surveillance as many people tend to think about it where it's like oh someone you know the state doesn't like someone and then they start looking at
what they're doing and they kind of like you know send undercovers or whatever to their house and so on um there's a slight difference between kind of that which is i think a common conception of surveillance and then what happens with mass surveillance and algorithmic surveillance where in actual fact the behavior of everybody is gathered and scraped together in order to then be analyzed and then surface who is you know doing something abnormal and who should the state actually be looking at and so and i and i want to make that point and i think it's very important to make that point because um uh so often you know when when we talk about privacy and surveillance people are like oh well if i haven't done anything wrong you know why would this state come after me or so on and so forth but when i when actually you know
all of your digital activities online you know interact with everybody else's digital activities in a way that might harm others so when you're being surveilled it's not necessarily that will harm you but it will harm other people because you're part of setting a norm and thereby kind of like surfacing abnormal behavior or becoming a target um in terms of you know everything from kind of credit rating to terrorism lists so you know what we're talking about here is really like what mass surveillance does at a kind of larger scale um and at a kind of at the scale of society more generally and you know how that curbs the the freedoms of society more generally yeah like one uh we did this of year review episodes a few weeks ago and uh
one of the things like i thought mentioned there like i've been thinking about a little bit about so if you if you think about crypto as this you know technology that you know threatens the you know existing powers you know like the nation states the uh banks and and then if if you want to kind of suppress that right as a nation said no like how how would you do that i mean it seems like one thing is if you look at the tax rules around crypto it's like basically very confusing and nebulous and unclear for the most part and then i think anyone who's now going to do like any defy nft stuff at scale is basically going to
be like unable to i think file their taxes properly right because it's just super hard so i i think that is actually you know in a way then it it will give this great leverage right because if you if you basically if everyone by default becomes like someone who you can potentially pin for because like oh you did this thing at that point and like was it properly filed well no so that like that's one one way like i could see this becoming like a big topic at some point you know what are your thoughts on that yeah i could go quickly on that i mean so we have uh a nim because we're a surveillance company
coming up hopes to fight surveillance that we uh we realize that there's a lot of legal issues actually uh ends up running a mixnet or vpn or tour like system is perfectly legal but what is definitely not clear what's definitely gray zone is a lot of money transfers right so you have a giant list of countries and people which are on america's blacklist that's the ofac list you've got politically exposed persons uh you have also different lists per country different rules per country and blockchain technology uh it has two issues that make it a gold mine for mass surveillance uh one is everything's over the internet so everything's naturally cross jurisdictional uh and second everyone self surveils by putting all
your transactions on the blockchain so we can prove that they happen and that blockchain is a very useful uh source of data and then combine that with mass surveillance when which is where most of data of any kind is being sent even off-chain data let's say lightning uh transactions and whatnot um and and you have an incredible source of data for any kind of uh any kind of legal prosecution and you remember you often sometimes you know you may not be prosecuted for what you think you're going to be prosecuted for it should be remembered for example that you know they couldn't take al capone out from part of the mafia but i think they got them on taxes if i remember correctly and that's what you call parallel construction argument those are actually
we believe quite common there's some evidence our nim's lawyer has actually used to pursue some of those cases around he brought the first cases up to the u.s court about mass surveillance uh being used for parallel construction arguments in the case of guantanamo detainees right so they'll grab some random you know shepherd from afghanistan or something and say hey you're a part of al-qaeda and of course he was actually just a shepherd but they could kind of go through everyone's phone records and a cousin of a cousin of his brother's friends somehow communicated to someone who was on a blacklist and that could happen to anyone not just uh that could happen in the united states that can happen anywhere anytime and i think these kinds of uh going off which i said these mass surveillance combined with
blockchain technology combined with lack of legal certainty and regulatory uncertainty are our gold mine for repression and i honestly look i'm an american i used to be at mit i worked with gary i don't think that they are going to clarify these rules because why would you if you clarify the rules about mass surveillance you clarify the rules about currencies uh that you have you lose ability to selectively prosecute people you don't like and what nation state in their right mind would give that up i sure wouldn't i mean and um so i i think the situation's actually gonna get worse i feel like we're actually maybe a calm before the storm particularly in terms of defy where i think the us government has kind
of said we don't care how many dumb foundations you have or weird uh dows or you think you're anonymous we're to pierce all through all that and we're going to make sure all the developers get adequately uh punished and that's uh that's a so we're entering a possibly really dangerous peer in history for cryptocurrency uh probably the most dangerous and that's why i think technology like nim is really needed so jaya i read you i read a blog post uh pen by you uh in which you cite the cypherpunk manifesto um in that um transparencies for for the powerful um privacies for the rest of us so basically the the second part kind of
becomes super clear um in the context of what you and harry just talked about um can you talk about the first part transparency for the powerful and how is how how should we basically what kind of transparency do we talk about here and uh who should count as powerful um yeah i think that's a that's a really good question like it's it's relatively easy to answer that question when we're talking about kind of traditional institutional um and governance frameworks where you have a kind of clear division between a public and a private sector right where it's like okay people who hold uh you know public office people who make decisions that affect like broad uh
parts of the population they need to be held accountable for those decisions and for the actions that they're taking right whether that's in public office or whether they work for a big company or you know whatever so it's like that's it's a pretty kind of easy uh topic to explain when we have kind of relatively clear power structures that are institutionalized um what i wrote that blog post about however was how to address that question when it comes to decentralized networks and what i've found is that many times people's understanding of surveillance and privacy tends to be very crude so they tend to think like either you know we have privacy and then everything is private and anybody can do whatever they want um or we have like complete surveillance right
um when in actual fact that's not how uh infrastructures and architectures actually work um and that's not really how we would want them to work because you know if there is a kind of like blanket privacy then you know you you very quickly tend to see uh power grabs that then remain invisible and people who have inordinate decision-making power that affect other people and that can no longer be held accountable so how do you deal with that in supposedly decentralized infrastructures and architectures um and so here the kind of argument that i put forward in that blog post is you know who is the powerful in this in this case when we're talking about a decentralized network um well it is the it is the protocol and the network itself right the or the technical organization of the
network and so therefore you know people that are running uh core parts of that infrastructure need to be able to you know need to be known and needs to be able to help be held accountable and so in our case we're talking about um node runners and this this extends a little bit further and i don't mean node runners in terms of the private people that run you know behind running the notes but i'm talking about the actual nodes themselves so and this extends to also questions of information security so um i've been having long conversations with claudia diaz our chief scientist at nim who designs a lot of the both the kind of token economics and the uh the security side of the infrastructure and of the mix that and
she talks about this a lot too and she she wrote a paper um that was looking at how to verify the location of nodes um and you know we were very quickly conscious of the fact that like for for for a lot of people who think about you know privacy and surveillance you know the idea of being able to verify the ip address and or the exact kind of geolocation of a node sounds like you know sounds like surveillance it sounds like a kind of like uh you know we're we're we're spying on the whole network or whatever but actually like when you think about it in this time this kind of term we need to know the location of the nodes in order to understand that the network is sufficiently decentralized across geographies in order to be able to assess the security parameters of the
mixnet and know that nodes are not colluding right um and uh and we need to do that in order to secure the privacy of the end user and so here the kind of you know i hope that i'm making myself clear that fundamentally it's it's a kind of extension of the original cypherpunk saying of transparency for the powerful privacy for the rest of us but an extension of that into a context of where we where we're operating with decentralized networks yeah i'll just add a little i'll jump in here really quickly i mean it's very important to remember that the purpose of nim is to provision anonymous communication and privacy to end users that's that's the goal and a lot of uh our goal is not to make
uh um node runners themselves mix nodes hide their ip address they're doing the work that's rewarded to help we may discuss what a mixnet is uh soon to help users get get get uh get privacy and the people that have the powers networks and we're seeing this you know dows with developers increased transparency and governance uh protocol changes and upgrades these are all things that you know i'm very interested in a little bit skeptical in some dials but overall i think the the moot at least that the thought the heart is in the right place these are things which are necessary in order to build a real privacy enhancing network so let's take the example of tor or a lot of other networks where you
know you have a lot of users in the us all the node runners in germany so you're just shuffling packets between the us and germany then back to the us where all of silicon valley is uh that's not very that's not very great wouldn't it be wonderful or better if we could guarantee that your for example as was done early versions of wikileaks that uh your packets went through multiple jurisdictions that would be great but we can't do that unless we know we don't want to know the identity of the node operator we can't do that unless we know where their actual ip address is and where their machine is so you know we could sort of say we could guarantee uh that for example you know a a a a piece of traffic that's going
that's very important goes through not just german the us but also for example could go through iceland or could go through switzerland and uh that's really important to do in a very in a very practical sense likewise we also need transparency and we need thing other technology that increases transparency like reproducible builds in order that users can trust our technology and that if i get uh you know rubber hose attack i get kidnapped jaya gets kidnapped or one of our developers goes goes rogue and decides to put a back door because you guys out there the end users can catch it that's again another good reason to push open source and then make sure that no matter what happens the network is fully functional
and there's no dependencies on the team or and we have enough technical safeguards that no group of node runners can control and corrupt the network maybe this is like a little bit of a detour but just something that came up that was like i thought was very interesting so there's this guy i don't know if you remember him but it's james d'angelo he made this like really fantastic bitcoin videos like a long long time ago like eight years ago or something seven years ago but i remember one of the things that he kind of he then sort of stopped working on bitcoin and he started this research basically on like transparency and his argument that he was making was that actually
transparency was something that was used to corrupt a lot like u.s politics right where you basically have like you know in the senate or in the congress you know people would like vote a certain way and then you'd have all of these organizations you'd be like okay no we need to you know these interest groups who then would basically give donations based on like some voting records and we're basically able to like use that data effectively to sort of corrupt that process i'm curious if you have any like thoughts on this idea i mean i don't know the specifics of that story um but i do think that there is a lot that still needs to that needs to be worked out in terms of how the
dynamics around transparency and privacy operate especially going forward into the kind of web 3 space um harry kind of touched on it a little bit earlier like there is some uh let's say there's some kind of you know industry standards that have kind that have kind of emerged around this stuff so like it's fairly kind of uh commonly agreed upon that you know code needs to be transparent right open source is kind of like is pretty important people understand the security reasons for that um and there's also kind of accountability reasons for that right it's like how is this thing actually uh you know how how does this thing how you know what does the code look like of this you know you know key infrastructure that i'm actually depending on um so there's certain kinds of let's say
norms and standards and and um ways of doing things that that are emerging in and around the web3 space but on the question of kind of transparent ledgers and on um you know node operators and all this kind of stuff there's a lot of like murky stuff that hasn't quite been worked out yet you know where is the line between surveillance and and privacy um where is the line between surveillance and accountability um you know it's kind of this is stuff that that needs to be really kind of thought through in some detail and i think this this coming i think this coming year i mean with the kind of like boom of web 3 that we've been seeing recently i think with this coming year we're going to see privacy really become kind of one of the core
uh parts of the conversation one of the kind of core topics in in the debate and back to the government issue i mean i i find it hilarious so i'm visiting the united states uh and all of my uh friends here in the uh wonderful world of bitcoin and blockchain technologies are really talking about how they really are going to have to lobby the us government and uh it i uh you know we're not going to touch the u.s that's our name doesn't touch the u.s that's one of our primary objectives actually in existence however you know i said well how are you going to lobby the us government and they said oh we're just going to like you know pay campaign donations to a group of random senators and
representatives i was like wow and i was discussing this actually with my friends in the middle east uh they said you know in the middle east the us media would just call that corruption uh you know they they well you you don't you know if i say hey would it be great if you passed this law here's 50k uh and this law saves me millions or you know half billions of dollars that that's that's corruption that's not really uh i mean lobbying is a very nice sounding word but that's not and i i honestly don't think i mean maybe i i've lived in europe too long but i don't think that happens in europe at least not to the same extent i see it in the u.s giant can remind me but i've never heard of this kind of lobbying at least in switzerland or france which is
where i've primarily been i'm pretty sure it's illegal um but it's normalized the u.s and the u.s has some virtue signaling around it would say oh it's just lobbying and i'm glad you know to be honest if i were a u.s blockchain company i would be very worried i'd be doing it too uh but that being said uh the fact that governments themselves can be so heavily influenced by uh existing wealthy powers and of course obviously the traditional banking sector is probably one of the most wealthiest uh uh powers and imaginable i think is think bodes very poorly for the future of the united states uh in general i don't see a difference uh between the
u.s government uh how they deal with lobbying corruption in uh what they would call uh some third world third-rate government i think it's basically the same um and i i hope at some point the us government actually gets its act together and purges all this corruption out of the system as a as a european i i can assure you similar things happen in europe um and i mean usually it's not fifty thousand dollars against a piece of legislation's legislation but it's um it's mostly um a little bit more um well less overt uh so basically it's kind of like ingratiating yourself and basically um
giving people talking points so they can look knowledgeable and so on and but it absolutely happens here i think it happens everywhere in the world um but uh i mean zooming out a bit again it seems there is a very dedicated group of people who are whose aim it is um to protect our privacy online um but to me this actually starkly contrasts with what probably 98 of internet users actually act like right so almost everyone doesn't give a second though to the data trailer they leave behind on the internet so do you think
that this is a um french movement kind of like the climate change activism you were involved in in the past that's gonna make it to the mainstream or do you think this is something do you think people will start to care that they give their data to google and facebook and um you know all the other large corporations that mine it um or do you think this will have to this this will stay um the this will stay like a pet peeve for a small vocal group of privacy defenders um i'm happy to take this harry so
i think we can see from the downloads of signal app last year that like people absolutely do care um the problem is not whether people care the problem is that it's actually really difficult to defend your privacy online as an individual user in the current uh infrastructure setup and so what nim you know aims to do is to really kind of protect uh communications at the network layer right so we're not kind of like gonna provide yet another tool to a user who feels a little bit paranoid to try and kind of patch up a leaking system is really to try and kind of address some of the problems at the kind of deep infrastructure layer so that's the kind of you know that's how nim positions um ourselves and which also means that you
know we would we would sit you know the mixed mixnet would sit under applications like signal and other kind of privacy um applications um i do not think that privacy is gonna remain a pet peeve i think actually it's gonna be one of the main topics going forward in the next year and for some um really important reasons uh first of all you know the covet pandemic has been going on for for a couple of years now which means that pretty much every single activity that we do as humans currently now takes place by a digital infrastructures and that includes you know school shopping um you know all the way through to probably sex and and everything in between right so
um what that means is that you know we actually are more exposed to surveillance than we have ever been before and i think that's going to start to dawn on people um so that's kind of on the bigger scale and then i think some of the kind of as i mentioned earlier i think some of the privacy issues with web 3 are really going to start to come to the foreground um we've seen people like moxie already write posts about this um and you know there's more and more debates in the web3 space around privacy and then uh and then finally also you know payments you know cash is disappearing almost all payments are becoming digital um so the question of kind of economic and financial surveillance is going to start to become more and more uh apparent to people as well as they start to see you
know everything from their credit records to you know who knows what you know start to be affected by um economic surveillance and it's something that you know we can even talk about this across different types of um bodies right so there one thing is like whether the individual cares and i think the individual absolutely does care you can see that in the amount of netflix tv series that are coming out about um facebook and various other kind of big tech surveillance driven platforms um but i think you know uh also more importantly companies and organizations are really going to start to care um you know corporate surveillance and so on um and all the way through to to governments you know like uh the conundrum that governments have is basically like how to balance
you know they're you know keeping you know national communication infrastructure secure but then governments are always trying to sneak in a back door which actually like up their security anyway um and so you know what we really need is a solution that cuts across all these diverse actors and that do not have that kind of like the problem of of having some kind of inherent interest in surveilling others which is the problem of governments and the problem of corporations um and that's why nem is decentralized and that's why you know nimr is trying to operate at the infrastructure layer rather than just um offering yet another patch to a leaking system for for an end user so let's talk about nim so let's talk about the infrastructure that you guys provide um can you maybe um
describe in a nutshell what nim sets out to do and how it sets out to do that and then we can kind of delve deeper into the details yeah sure we could essentially what imagine that you have an entities an adversary an enemy that is watching your every move it's a bit like that 80s song right uh every move you make on the internet they're watching every packet that comes in and out the network how would you defend against that so you could say oh don't worry we'll just use the vpn but you know if the adversary can watch all the packets coming in the vpn and all the packets coming out the vpn so
they essentially get a kind of fingerprint not just with the ip address but the timing and the volume the kind of pattern of those packets so vpns don't work uh and you know tor is much much better than a vpn but you can kind of run the same attack on tour you can say look at all the packets coming in the torrent you know look at all packs coming out the exit node and we can do some machine learning on those and and figure out what stream of packets is coming from which user so then the question is how would you defend against someone watching who's watching all your packets and who's watching everyone you might be talking to his packets how would you defend against such a powerful enemy and the answer is actually pretty intuitive which is you have just like tor you have
nodes in between you give your packets to someone else other people give those packets to someone else that person if you imagine the packets are like a deck of cards they mix the packets up that's why it's called a mixnet you know you could imagine it like shuffling a deck of cards if you're a mathematician a random permutation and then they release those packets in our case using a statistical process called a poisson process other people have different takes on it and those pack those packets are then mixed up so they're not coming in and out the node in the same order then they go to other nodes and they continue to mix those packets up and after a little bit of mixing even an adversary which is watching the entire network with this god's eye view
can't figure out who sent what packet to what person and so that is honestly currently the most scalable uh defense against master fans we have today and that's what nim is nim is a mixnet it's composed of servers nodes which mix packets on behalf of users now the big question with mixnets and mixnets are not new they predate bitcoin adam back was commercialized with zero knowledge systems before he founded block stream um cypherpunk's probably such a uh were using mixed early mix nets like mixmaster and makes it minion to disguise their email so they've been around a while but the problem
uh that afflicts mix that says where do you get uh all these people how do you get those people to run nodes that's a really good question uh because you're assuming a decentralized network uh the security and the privacy is dependent on having multiple independent parties uh use the uh running these nodes and so obviously we can take the lessons that we've learned from bitcoin and apply those to mixed networks we can say well can we reward mixed networks using a kind of reputation token to say look this mix node's really doing its job mixing this other guy he's just dropping all of his packets and that's what nim does so nim takes a mixed network design particular design called lupix built by george and anya who are both now at nim
and uh we improve upon it and one way we improve upon it is we add an incentive layer on top and a rep which is built on our essentially a reputation tracking mechanism and that is the provision uh this kind of privacy that makes that uh mix that's give you with also high quality of service and the mix that's kind of fast and reliable and people that can't run mixed nodes very well eventually carry less and less traffic over time and those that uh do mix traffic correctly and fast get more and more rewards and we think that's uh do you think this is actually quite a large breakthrough this system has not quite been developed before and although there's some other things options in the space and we
really look forward to launching it at main net pretty soon so can you also talk a little bit about you know like the use cases like how like once name is live and functions well like what are some of the ways that you see people interacting and using interacting with and using them i mean in the ideal world you wouldn't have to know about it because your wallet for example your you know monero wallet your zcash wallet your bitcoin wallet your liquid bitcoin wallet probably for some privacy on chain would know about them it would just run them automatically just as when you download tor browser you're using tor automatically now obviously uh that's not the case right now
but we are going to build infrastructure such that people can easily plug in currently if you're a power user nim uses the same kind of sox5 interface as tor you can just go into key base or firefox and kind of plug in you can start a nim client on your machine and plug that local gateway ips address into your configuration you can run apps through nem through the testnet right now in the future uh we'd like to see a number of applications actually natively integrate nem i've discussed it with moxie from signal of course he's very skeptical in the decentralization aspects uh you know we could imagine i think instant messaging is a great use case for them uh
surprisingly uh you know there's been previous work which seems to point to the fact that things were a few packets being dropped or kind of okay would also be good for mixed nets so this would include voice and audio uh voi over the internet is pretty good maybe even video over the internet things that can buffer uh things that uh things that aren't very good for nim it's probably web browsing i think tor is probably the best option you have right now i don't really see that changing uh particularly anytime particularly soon uh so i use tor every day and and tour is more optimized for things like web browsing where you have large amounts of pockets packets sorry uh being streamed through in a synchronous manner and mixed that's
because this mixing process it's a little bit more latency and it's better for apps that have a more asynchronous message based approach to traffic so hopefully this stuff's just built in by default uh we'll let users use it and add it to apps as they see fit just as they can currently do with tor and we eventually hope that privacy enhanced apps take network privacy seriously currently to our knowledge almost no one does accept tor and that they build uh them directly into their applications so as i understand it there are two key improvements here right so basically one um is the fact that um every package is independently routed and thus reordered
kind of making it harder to uh to figure out that they kind of belong together and the other thing is um that um there's an incentive layer added right we there's one counterintuitive point in so far is that we allow you to add fake traffic as well so let's say none of people use the network you still want your packets to be anonymous we can basically add a fake cover dummy whatever you want to call it traffic into the system to guarantee certain levels of nominees even there's not enough users now the great thing with nem is thanks to the way that we built our network statistically the more people that use the network counter intuitively the faster the network gets the less
covered traffic you need if you want to be anonymous there's a giant crowd of people using the network then you don't need that much covered traffic you can actually mix those packets for less amount of time uh while maintaining the same amount of privacy that's not the case uh so we kind of with blockchain systems we're essentially the way to think about blockchain systems it's kind of an auction for a finite amount of of of a block space so we kind of scale like web servers you know more traffic comes in great we can add more mixed nodes and i'll let uh jaya explain the incentive system here in a second but i do want to point out that it's not just independently rounding and mixing the packets also adding fake traffic and balancing that fake traffic and the amount of mixing so we can achieve
scalability and high quality of service charlie do you want to just take over on the incentive stuff real quick um yeah sure i'm just trying to kind of think looping back on uh frederika's um question on like the improvements i guess um yeah we've been kind of like laying out a lot of the kind of like technical details but i wanted to kind of make it a little bit maybe a bit more clear for kind of non-technical audiences um what the improvement really is here um so you know it's almost like you got to think of kind of like you know the the stack in layers and so where you might have a wallet or a messaging application and so on that might have kind of like end-to-end encryption so you know we're
seeing this a lot and like also to speak to your earlier question of like whether people actually care you know once you start seeing like what's up try start to advertise that they're privacy friendly they have end and encryption so on then you know that okay this is something that people are aware of privacy is something that actually people care about but what's important to understand is that even when the messages is encrypted um the actual the actual metadata and the patterns of communications are still visible and that can actually reveal a lot more than people normally think so it can reveal you know where you're you know where you're located who you're speaking to how often you're speaking to them what time of the day which can reveal a lot about your relationship to that person so say you're sending a message to to
one person at 11 o'clock at night every single day then that's very different type of relationship than if you're sending a message to someone you know um at 10 in the morning or whatever so it's like there's a lot of uh important information that um is revealed uh and that is in fact as uh claudia diaz also likes to say is in fact a lot more kind of machine readable um that gets revealed you know rather than the content of the message and so what nim does is it adds that kind of you know a deeper layer of protection to communications that does not allow for um surveillance of the patterns of communication and the metadata of communication so that's what the that's the kind of major improvement to existing um private messaging apps or or
wallets and so on that makes a lot of sense um you said that nim can be used as a service for all kinds of applications right so basically blockchain and non-blockchain based so can you maybe um go into how nim works together with other other applications so we see what's the interface like yeah so i mean i think i i kind of leaned towards that earlier you know nim was originally not built with blockchain applications in mind it was built with people that wanted including european governments that wanted to avoid nsa surveillance in mind um and so the way you want to think about it
just as any application can run over vpn and if you're clever enough any application run over tor any application internet application can have its packets mixed by nem any application can run over nem in detail how that works is that just like tor or vpn you have to install a local client that client has a library that kind of starts that process out for you that connects to the mixnet that application your application that you want to use them has to somehow talk to that library currently there's a standardized protocol that's also used by vpns and tor called sox5 and we just support that protocol the box so in a lot of applications there's a way to kind of natively run your traffic through sox5 proxies and that's so things that were
built without nem even knowing about nem again key base firefox and these uh these sort of apps come to mind can easily plug into nem today which is quite nice so you know it's not a it's not just a dream in a white paper it's actually a real working code thanks to our excellent dev team who's been working around the clock for three years on this and so that being said there's nothing to prevent an application developer from what we call natively binding to them so that means they can actually just sort of add that add that library into their application bundle it with their application so you don't even think about it when you uh when you install uh when you install that application and you know in order because we imagine a quite
diverse number of applications wanting network level privacy this you know not just blockchain apps but everyone we've also built in a truly diverse diverse amount of ways to pay for them uh so nothing is free in this world but we try to make name as free as possible and allowed people to be very flexible so for example you know no one's going to pay for instant messaging that's crazy i definitely want it uh but we can imagine that instant messaging applications uh you know telegram signal and whatnot that are interested in them can just sort of you know use it in bulk for all of their users and they essentially uh just kind of purchase some bandwidth on them uh in bulk for all their users at once and then they use for the user it's just
free uh people that use bitcoin uh power users might want to for example you know whenever you use bitcoin you have a little transaction fee uh you can imagine a sort of similar model being approached with them where you you pay a little privacy fee to keep your network level data private and you know honestly we've built a very open-ended system in this regard so that not only can many different apps plug into them in very different ways including native bindings and proxy bindings and standardized bindings so it's really up to the developer or the even the power user how they want to plug into them we've also made it super uh easy for people to pay to pay for them in whatever way they want so i don't want
people to have to do something sketchy with an exchange just to use them i think that's crazy and that's blocking a lot of quote unquote web3 projects if i'm using them with a bitcoin wallet i want to pay for nem using just a little bit bitcoin and you know maybe an api call i think that's how most people who use bitcoin wallets for example interact with them let's talk about the business side um in a little bit um i i'm still not 100 sure on how it works so maybe if i give an example so basically say um i have um 500 packet packets that i want to send to brian right so basically um so i would um whom would i so basically would my application that i wanted to send
this out of already kind of distribute this to different nodes or do i have to send it to an initial node first and they kind of split it up or um because if i have to send everything to an initial node um then it's it's immediately clear that all of those come from me right so at least that node would know and if you surveyed all traffic um you could also know yeah so how nim works is that i download some client software it's important when you send traffic through them that each packet is encrypted and the same size and looks just like every other packet otherwise you could identify them so these packets are essentially uh your whatever let's say you're five three packets they're all made into let's say let's say they're
made into a thousand sphinx packets which is this format that we use to make all the packets identical that's done locally that's done on your machine by a client of what's called the nim client now the name client then wants to talk to the nem mix network now there's two ways to do that the first hop in the network which is obviously the most vulnerable is called the gateway so you could go to public gateway which could be just like a vpn provider it could be the app could have the gateway built in or if you're pretty paranoid you don't want anyone to know your your ip you can run your own gateway you can run it on a raspberry pi or whatever in your house on your you know on your machine your desktop machine probably not something you want to turn on an awful lot and
those packets go to the first hop into the network which is called the gateway uh that gateway then asks a blockchain a directory authority system for where's all the mixed nodes uh you know and that gets given to the client software the client software each time it's building the packet also builds a independent path for that packet through the network and that packet that path is not revealed it's encrypted in such a way that it only the next hop just like tour is revealed at every hop so you can imagine you have 500 packets you build a thousand sphinx packets inside of each things packet is a encrypted route those packets are then shipped with some
sort of timing delay which is again statistically chosen via fair process uh to a gateway that then ships it to the next hop in the path and then hop is to a mixed node and it goes through one mix node it goes through two mixes currently goes through three mixed nodes and then it comes out and at the last hop because again you told the packets when you were making them where they're going they know where they're going it's just the people that can view them can't see where they're going the last hop kind of just like in a taking the peels off of onion and onion encryption takes that final layer of encryption off it sees the address the ip address of where that packet should be going that
packet sent to the gateway or the exit node you can think of it that goes to that application or bob or whoever it is you're sending the packet to so again nodes are bundled locally they're all made identical they're given their path locally in an unbiasable and fair way using verifiable random functions another kind of cool crypto sent through the network when they go through the network then mix those themselves never they only know they don't know who sent it they don't know where it's going they only see the top it came from the next hop that's why you go through three mix nodes and that's why you go through five hops in general so you have a local client since the gateway three mixed nodes final gateway then to bob or
whoever your application is at the end and you think this would be very slow but actually you can get good anonymity by just a few seconds you know definitely minutes but also seconds through this process this process can be done pretty quickly thanks to the wonders of high speed modern cryptography is that more clear there's a bit more detail than i was expecting to give on a podcast but yeah okay great so just one question here like i'm because you you basically like mixing up the order potentially right now i think if you send something like a crypto transaction like okay this makes like obviously make sense to me but i wonder you know like are there particular types i mean you mentioned before i think like web browsing is something that wouldn't
wouldn't work well is that correct or like you know what are the types of application that like this works for and like what are the things where basically like this isn't suitable for a lot of people talk about nim versus tor but i do think you know there are certain applications where mixnets add too much latency and shipping packets in out of order don't really make sense so things like web browsing for example i think tor is a great solution for web browsing um however you remember that this gateway essentially is taking the sphinx packets and buffering them and reassembling them so even if the packets come in and out of order you know you're seeing enough packets to they kind of get buffered and decrypted and sent to the client that sorry that's actually done on the client
not the gateway and that that's uh that makes them suitable for certain kinds of applications cryptocurrency as you mentioned earlier instant messaging but also things that you that that where it's okay for things to be a bit slow in the buffer or for occasional packets to get lost i can do an audio or video call with you and some packets can get dropped uh you lose some of the quality but you can still do the call and i think mixedness would work for that um also you know i may be uploading or downloading a giant file obviously it seems going to be a little bit slow to use it with a mixnet but again i might not need to upload download that file in a few seconds i can wait a few minutes for that file to be chunked up and sent through a mixnet
uh that being said uh web browsing really is the hard problem and i think for the foreseeable future uh it will uh tour will be better for web browsing it's not theoretically impossible to send web browsing traffic to a mix that in fact we have uh we've done it you can we have instructions on how to do it on our website if you're interested in doing with firefox it's just that it's it's kind of slow and it's not i think the ideal it's not an ideal mechanism one thing to note with nim which is sort of interesting i actually forgot to mention earlier is that different applications will have different kind of there's kind of a trade option how fast you want to go and how anonymous you want to be and we actually let the user in the application choose that level we'll have some defaults reinforce but
you know you can say i don't really care about that this is like a this is a super private bitcoin transaction so i want this to take like i'm okay if it takes like a day to go through because it's really important that remains private and we can support that but we can also sort of say hey i'm just i want some privacy but i'm also would like to do an audio or video call with epicenter and this isn't super it's going to be public so it's not super important that i keep this 100 anonymous and we also will support that kind of application as well but i think what's also uh and what's fun about that point is that those different applications you know all of them together in the same mixnet actually enhances the security of each other because of the timing obfuscation
so it's you know the multiple different types of um application use is actually to the benefit of everyone oh yeah absolutely um can we talk about the incentivization that goes on top so basically if you if you look at traditional um privacy um systems like tor um as far as i understand that their tour is um is donation based and uh nim kind of has this in-built incentivization layer right so basically there's a token and everything um can you explain how that works um yes i'll be happy to explain how that works uh yeah so we have a utility token in nim which is called nim and at a very basic level um
the system is designed to incentivize mixed nodes to do the work of mixing data packets and not only to do the work but to do it well so the intention is to incentivize good quality of service rather than just um mixing so the way that that happens is that um the you know the system is kind of like quite kind of uniquely uh democratic in a sense because the performance of each mixed node is measured in the reputation in the community reputation so how that works is that anybody who holds nim tokens can actually delegate stake to a mix node um that they believe is gonna are gonna
perform good work a good work of mixing and if they do perform good work of mixing that mixed node will get rewarded in nim and then the people who have delegated to that mixed node will receive a share of that reward um so the way that the incentives works in this case then is that people who hold them are incentivized to delegate to nodes that perform well um nodes are incentivized to perform well because then they'll receive more delegations and the more delegations that they receive the more likely they are to be um selected to continue doing the mixing for the network so um the kind of that's a kind of like brief sketch a brief kind of overview of the reward system um but what's nice about it is that you know
you can have different levels of um expertise come in here you know where those running nodes might have you know a bit more kind of expertise in how to kind of set up a well-functioning node um but in actual fact anybody who holds nim tokens can participate in ensuring that the network operates well and and this is uh right this is like a cosmos sdk network guys is my understanding no so it is uh does it mean basically like all of the mixed node operators are also validators of the proof-of-stake network or are these like decoupled functions so these are decoupled functions so essentially you have to kind of think about it as like
um almost almost two separate systems i mean that are very integrated but um kind of two separate systems so the mixnet with the mixed nodes you know operate the mixnet um and uh and then separately to that you know and the mixnet makes use of a cosmos blockchain um which is then operated by a set of validators um the set of validators currently is a permission set but we're going to be de-permissioning later this year and it's a permission set because it is a core aspect to the infrastructure so we're trying to kind of like match that to to real skin in the game in the real world [Music] is is there a cap on how much traffic can go through any individual node so basically if if a node is a fantastic
mixer and always available um it is a way to guard against um emerging centralization yeah so in actual fact um there is a certain amount there's a certain number of mixed nodes um that need to be present in the mixnet for every time period every epoch so to speak um and and that's calculated partially on the basis of how many um bandwidth credentials have been purchased with nim tokens so it's a kind of measure of what is the demand going to be like for the next epoch right um which will then calculate how many mixed nodes do we need and those mixed nodes are distributed across three layers um there is an equal amount of traffic
going through every mix node um per epoch basically so in that sense there there won't ever be centralization um and there's another element to the to counsel decentralization which means that um essentially your you know the the total amount of delegations that you receive as a node you know will increase the likelihood of your selection for the next round but only up to a certain point so there's what's called a soft cap um after which you know it won't make much difference if like you know 20 000 more people delegate to you um yeah so there's there's inbuilt kind of let's say resistance to centralization and what's the state of this today so i
i understand you have a test net life right yeah so we um just before everyone broke up for the holiday we launched our permanent test net called sandbox which mimics all the kind of mainnet smart contracts and incentive schemes and yeah and people are more than welcome to to get involved in and try it out it's very important for us to to kind of give to allow people to to kind of try out running nodes and test out the reward systems and try building things on nim before they actually kind of get involved in the mainnet so we welcome everyone to to have a look at sandbox final topic we wanted to ask about uh so
you know you were hired you were on the podcast before where the topic was like about julian assange and and actually we had we did a podcast uh just before the holidays with amir taki and uh some of his co-founders and then we talked a little bit about about the two and he mentioned that there was some sort of assange dao so i was just curious like what's what's the update on this is still something like you're involved in and you know like what's going on there oh i can i can give a quick update on that yeah so i'm uh uh old friends with uh
uh julian and uh eventually his family um actually there's a there's a quite funny story when i was having all my problems uh due to climate change activism i i went to see the wikileaks launch i think it was like 2008 or nine at a chaos computer congress that was still in berlin and uh i said to uh the wikileaks folks i said look uh you're all gonna go to jail and i've never one of them said uh you know we're not dissidents we just create tools for distance i said i don't think the fbi is gonna see it that way um and i believe i was right and then uh you know i felt quite bad about uh you know um no one is perfect uh obviously julian's not perfect i'm certainly not perfect
uh but therefore you know no one particularly for releasing truthful information deserves to be imprisoned and tortured so you know i visited julian embassy and we would discuss lots of topics to be honest he was pretty depressed sometimes and i would say well look this blockchain stuff you know these these people from bitcoin uh uh you know the cryptokitties people whatever you want to call it they're actually doing really amazing stuff there's a new a new social movement happening here and honestly he was a little bit skeptical but he got into it uh and uh you know when when they finally um i was uh i visited the embassy i think it was the last person on the guest list uh before they closed the guest list and
then two or three months later they busted and took him out put him in prison and so i always felt a little bit of a personal responsibility to julian now you know i mean we philosophically disagree on some points i don't believe information uh information may be necessary to help uh freedom but you know human freedom is is caused by human social movements and human people not by technology uh technology can help those social movements uh but nonetheless i think julian did a remarkable uh job uh and building some crucial parts of technology and releasing a number of very true uh true things ending a really bloody war in iraq and therefore i believe deserves to be supported uh and
you know it's really hard to support you know particularly in the cryptocurrency space you think that there would be tons of donations and whatnot but there kind of isn't uh you know most people that have bitcoin would like to hold it so they get more rich uh and so uh i went with julian's father to bitcoin conference miami uh i don't think julian's father was particularly successful in fundraising at that event but then we saw this crazy nft stuff happening where people were selling nfts for uh lots of of money uh and i was shocked i don't own any nfts i don't even own any ethereum so i don't uh to be honest it's a separate world for me i feel a bit
a bit old-fashioned but regardless uh i do believe that one of the best things that could happen uh for julian would be if he could raise enough money to pay his lawyers and get out and i believe uh that will hopefully happen and i believe the family is working on some sort of uh nft sale and of course just as we've seen in previous sales dalles have arisen uh to help julianne uh to help uh and to work on his uh kind of nft auction now you know i'm not part of the the dao uh i i just connected the dao uh the people who are in it with his uh
with his family because they're friends of mine and i i don't want to see uh julian rod in jail and i i do think that i i would like to regardless of what people think about uh wikileaks per se uh i think it's great but i understand if some people don't no one deserves to be imprisoned and die in jail for releasing information you know i had a previous friend of mine aaron schwartz driven to suicide over this when he released what was actually public information about uh about the us court system was called the pacer database and then later they kind of kind of blackmailed him over uh open access data which now we're seeing is super useful for the world in the form of libgen and i think we're gonna see
more and more of these cases and i would actually you know i mean regardless of how you feel about the details of each individual cases i would actually see like to see lots of dowels form uh for generically uh supporting a political prisoners particularly political presumed cryptography but also wider political prisoners who are fighting for human freedom so this is not just uh you know virgil griffith obviously uh was victimized by the u.s state i can imagine there'll be plenty more in the future and so i think i really want to support this kind of movement because to be honest we're all everyone is talking about funding public goods but one of the public goods that we all take
advantage of and we don't even think about is the fact that we are free we are not imprisoned in a room uh unless you're on a covered lockdown and uh and we are able to kind of talk to we want to when we want to and i think that fundamental freedom should be upheld and there will be a lot more court cases in the future so i would like to see the crypto community become more generous and fight for our common our common freedom cool well thanks so much harry and jaya i think that's a great place to wrap up and it was really a pleasure to yeah speak with you both about nim excited to see like where it's gonna go and i think it's yeah very needed definitely
so i hope it's going to be part of lots of crypto wallets and messaging apps soon yeah the way i think about nim uh before saying goodbye is that i think there are many things needed in crypto and the way you think about them is they're like each component is like a finger that can make a fist and that fist can kind of fight for our freedom and nim is just one very small finger uh but we think it's an important one as one that no one has built is one that should be built i do look forward to hearing what everyone else is building and i'll be i'll keep listening to your podcast so thanks a lot where can people come to to find out more about them and get involved um yeah so thanks very much for having us on the podcast and um i would very
much invite people to keep following um our project we will be we are have already slowly begun to launch the mixnet and take the steps to to towards mainnet which will come very very soon um so yeah you can follow us on you know telegram and twitter the usual channels um and of course our website so keep an eye cool thanks so much