Zcash & NEAR: Zooko & Illia on Privacy, Intents, and the Future

Captions as YouTube publishes them, usually machine-generated, so product and people names can be misheard.

1. 0:00

   so hello Z it's really great to chat um been a second since we've seen each other but yeah being T Bunch around this launch so uh yeah I mean I'm milia profounder of near um we're building a lot of different Technologies from across blockchain space from W itself to interoperability solution Chain Solutions as well as a new function so we be good to dilated all of this we introduce yourself as

2. 0:36

   well I'm Zuko I'm one of the founders of zcash many years ago and now I work at a nonprofit named shielded Labs that's one of the current support organizations for zcash awesome so we just had this launch um few days ago with uh zash and being added to near and 10 W the I mean the responsive Community being amazing I think the G commity very engaged also very like strongly pushing for privacy which is really exciting I think yeah all the near people as well

3. 1:14

   are like holy it's exciting to to pursue this B um B responds uh so yeah maybe it's like two tribes have met and it's exciting find out there's another tribe there yeah yeah um so maybe would be good to kind of from your side well I mean maybe you can like introduce how obviously Z Cas been involved I mean around for a long time I think like when I was starting Z Cas was already around um and how you how you see kind

4. 1:51

   of the zcash playing in this like multichain world uh and then we can talk about you know kind of what this int as well like some AI ideas just F I'm really excited about what I call in what we call interop like zcash interacting with other blockchains because I'm more of a patriate for the Ze asset than the zcash blockchain like the zcash blockchain is awesome as you mentioned it's a battle scarred veteran

5. 2:24

   it launched at the end of 2016 um so it's I think it's one of the few crypto projects from that era that are still relevant yeah still less and and kicking so yeah congrats to that by way yeah I'm like four years in and it's and it's you can feel every year yeah yeah like that those pictures of President Obama being all gray after being in the white house that's me and so zcash has all these great like sort of unique position and this great Community like you mentioned and one

6. 3:04

   thing it hasn't had is good trustless interop like until now if zcash want to interact with the rest of the crypto World either to go through a centralized exchange Litecoin base or they use one of those instant exchangers like let's exchange. where you just send them your money and then you count on them sending you back different coins soon um and I so I just played with the near intense integration um today and my

7. 3:37

   simplistic understanding of it is it's Pro currently in the current version of it it provides basically the same functionality as let exchange. except that instead of a centralized server on the other side it's a smart contract running on the near blockchain on the other side is that accurate yeah so I mean uh might be good explained architecture so we have kind of set of technologies that are powering this and they're all at different Readiness and like of maturity uh state

8. 4:08

   of maturity um so the intent kind of at the end right is is this idea that instead of prescribing how exactly there's exchange or any any actually State transition should happen you're just effectively asking someone to figure it out and give you B result right so for trading it's simple it's like Z I want the BDC but it also can be like way more complex it can be get me pizza at home right that's saying yeah like an

9. 4:40

   abstraction an abstraction Beyond like a a swap order or a limit order exactly yeah but but but like one kind of important note we also using this Beyond like trading and even Beyond blockchain cases like we think this is not appable across like whatever you put in Google that is affecting your intent as well right like I need a plumers or you know I want to run a marketing so so that's kind of the like I mean we're aiming

10. 5:11

    toward that Vision right now the indeed the swap is probably the simplest right and most verifiable kind of module of that to ship the simplest thing first yeah and so so so the way it works right when you say hey for example I have one near and I want some Z you're actually sending a kind of into this uh network of solvers you're sending a request saying hey do someone wants to fulfill me this trade right and so anyone can respond saying okay cool I'll take your

11. 5:45

    one year and I will give you whatever the market price is T and while doing this uh they actually sign on that commitment so they respond you back with the commitment to uh actually do this exchange so can I interrupt so that sounds like very similar to what I do with the centralized swappers like let's exchange doio it seems like the difference is uh like correct me if I'm wrong anyone

12. 6:21

    can run one of these solvers correct yeah anyone can join and marketplace right exactly yeah and so there solver like there's an example solver Cod already open source for people to you know play with this if you want to run your own Z you know Z to BC for example uh with Prett provision so this is like so when I did this swap earlier today I swapped some Z out of my shielded wallet in return for some eth U and the reason I picked eth

13. 6:53

    is because eth was already activated in my Brave wallet the brave web browser and that's that's where it was going to go so um you're saying that when I did that there was an Open Marketplace of solvers who to satisfy and whichever one want it got the uh like some transaction fees yeah I mean they've actually bid on whatever price they like they would want to satisfy your trade right which they

14. 7:23

    include their slipage and their effecting Fe in that right so then you actually get all the quotes right so you don't you I mean sometimes you get one but you may get actually multiple your your application without multiple quote and you application chooses the best price the cool thing here actually this Rec providers are bidding to get you to become their yeah and what's really good about it is that as a user I wasn't aware of any of that exactly I just use near Das

15. 7:56

    in.org or app. near- intense org and entered in a pretty simple UI that I wanted to send Z and receive e uh and then it showed me the price yeah accepted and then it worked yeah and so so that's kind of the mechanism on on on like for this kind of effectively it's in the offchain commitment power because like you effectively finding and committing to something offchain then it settles on chain and then when you want to deposit withdraw

16. 8:30

    we use this kind of chain signature technology which is a multi party computation of sub Val effec holding the Assets in different sh right so kind of in flight and especially for liquidity providers solvers the assets are held by this uh NPC providers now there's a caveat that's new networks because it takes some time to get uh NPC new networks uh this is like sometimes it done to to POA so there is like kind of

17. 9:02

    a transition period where first uh we like we added to POA and then we transition it to a full NPC Network and so that's kind of the process so the Z the Z that I sent went into a wall gcash spending key that was controlled by some proof of authority set of right right now yes and then what it will get upgraded is it will be a key that is held by kind of distributed NPC network

18. 9:33

    of Subs validators we're also putting that code itself into to as well to kind of secure this oblivious Ram to secure that even more so that's kind of the progression here like you know you start like maybe more more limited you know uh security but kind of we ramping it up and was more so the near intense Ze support is currently in the POA stage proof of authority stage what assets are the furthest along in this security upgrade

19. 10:05

    pipeline so so BDC s uh eum uh main net are all kind of in like effectly already supported by Omni Bridge um and so those are I mean Omni Bridge that's the name of the yeah so Omni bridge is like kind of higher security version of it yeah I mean it's effectively use the chain sign just to hold app kind of and like among other things also can bridge outsets between so you don't need to use intense but again this is like this is more of a

20. 10:38

    developer tooling that intent just uses everything controller right yeah which is thankfully hidden from the end user like me exactly yeah so that I mean that's a big kind of idea with the whole chain of traction we've been pushing is there's a lot of technologies that like the web three space is building which are extremely important but we we actually make them products like the bridge is a product in which is

21. 11:05

    like very strange like as a user I don't care about I care about using the app and just like things figur figured out on the backend yeah years ago I was really excited about like cloud computing like you could rent server VMS quickly on Amazon web services or whatever um but I always think of the current web 3 industry as being like as if you wanted to install install Snapchat on your phone so you could send selfies to your

22. 11:39

    friends but then like it offers you an AWS App instead you're like what is this I just wanted to send selfies to my friends so like once it's mature then my opinion is that approximately no users will be aware of what blockchain like nobody uses Snapchat knows whether it's using AWS or it's using a different service or data center or whatever if they knew that there would probably be something wrong so I think it'll be like that with

23. 12:10

    blockchains like people they want to send and receive money they want to invest they want to send selfies to their friends there's these things which I guess you call intense which is what the user actually wants to do and uh using ethereum or Sana or near or zcash as a blockchain is not one of those things things they want to do yeah so that's that's been the T since with choke uh and been pushing on and kind of that's why we like building a lot of this Tech under the hood but

24. 12:40

    trying to make UI as simple as possible uh including yeah so I mean I think the one dimension to like discuss is how can we both bring more of this kind of mindset to the cash but also how do we enable the cach users especially the ones who are storing their Assets in the cash right so like it's actually in Shield pools to make it easy for them to use other application in then kind of without like manually unshield SWAT brids you know like figure out how much

25. 13:14

    gas I need to pay on another chain and such for well I've had some pretty miserable experiences trying to use bridges in those ways where you like load up multiple apps and web pages and then you find out you don't have the right token to pay the gas halfway through yeah so I mean one of one of the ideas here like kind of with this intense is you can effectively say like hey I want to you whatever I want to buy something on

26. 13:44

    pumped F or I want to engage with uh I don't know hyper liquid or whatever openc magic and the idea that like I mean the same way solware is kind of find you like okay satisfy you this token they can also say hey I actually will execute this action on your behalf on that chain okay well I've heard about intense a lot and uh agoric orchestration and chain abstraction and n's I don't know

27. 14:14

    what's the word for the near thing where validator nodes have a te protected Shard of a spinning key what's that called we call yeah right that's cool too I've heard of all these things but now that you're telling you this now that we're talking I'm you you said you wondered about how to push the user interface in the right direction so it just made me realize the question is well so the user interface is all

28. 14:39

    controlled by the wallet or the app or the web browser what whatever app the user knows the name of it they know they launched the app on purpose and they trust the app now the question is uh how do the app and the user communicate to each other about the user intent yeah so I mean we have one really interesting project actually that doing this for Bitcoin so uh it's called satosi Port where you deposit Bitcoin into an address that is controlled by ch um and now hold on does that mean can

29. 15:18

    I interrupt you I have some simple-minded uh if there's anybody listening to this who finds this stuff as confusing as I do they will appreciate that I just interrupted say hold on you deposit so that means I have some Bitcoin and I want to use this app called Satoshi Port right now this app is a a smart contract right so it's there's no nobody who can steal my Bitcoin out of it correct a near smart contract it's a

30. 15:51

    near smart contract that s chain signatures as addresses on bitcoin and other chains right so can generate you a unique Bitcoin address where you deposit your Bitcoin and now what it allows you to do is you can actually in their app open applications on different chaines so you can open again Salon you can open hyper liquid you can open open t medium and you can do an action there it will translate that action into an intent

31. 16:25

    right like when you click it right it actually create a transaction on what kind of actions can I do in so for example you you want to like buy an nft onc so it actually they theyed open C directly into their application so you can open it inside they inject in terms of the user interface yeah yeah they name B I got to try this I've never bought an nft before but maybe I will now that I can do so starting with my Ze from my shielded pool so nobody will know it was me exactly and so so but like the interesting thing

32. 17:02

    here is that yeah so this allows to effectively under the hood do all the swapping funding even borrowing against Bitcoin so you don't even need to sell your Bitcoin you can actually borrow against it and then swap it into yeast and then fund a ethereum account and then buy an so the interesting thing is here with zcash is to actually do the same where the you keep the gcash you don't spend

33. 17:25

    it but oh that's interesting but does that mean if I'm also naive and simple-minded about Finance in addition to about technology so does that mean if the price of Bitcoin down yeah I mean you can get that mean the price yeah okay that's what I was getting that so then if the price of Bitcoin goes down while I'm doing that then I lose my Bitcoin and the thing that I was no the thing you oh okay but if the price of Bitcoin goes up in the meantime

34. 17:58

    then I get keep the thing that I bought against my Bitcoin to and the okay that makes sense add zcash support for that too so I can borrow against my Zach and buy nfts without letting go my Z for example yeah is Satoshi Port your project or is that another it's a team in the ecosystem that are building kind of on top of the technology on top of Technology they're listening should IDE yeah yeah the difference would be if you use zcash that you originally held

35. 18:38

    in your shielded wallet yeah then there'd be no link back to the source of your funds or your other transactions through through your Z exactly I like the idea of like holding on to it although I don't like the idea of getting liquidated and losing all set so well you know the saying that if the price goes down you just need to buy more can you do that in Satoshi port and like

36. 19:03

    put more in so you don't get liquidated I mean you can yes yeah can I register an intent to automatically buy more and put more in so that I don't get liquidated yes um that that is actually where I comes in uhuh really well so so we so we have the the other piece of technology in building is this te cloud which effectively allows to have uh I mean we call them agents but it's

37. 19:33

    literally just programs which can use AI which run inside te has effectively both Nerf viability and privacy and we actually can IOP you sure okay so I'm a user and there's this program which is an agent so it's just a program like a a rust program or something right and I want to delegate some of my property or my authority to this agent and because of the te then as long as they can't break the te

38. 20:16

    security then I know it will do whatever the rust program said it would do is that the mean thing and also it could have like a secret key or it could do some stuff and then not displ to the company that owns the machine with the te in it what it did so I can get some kind of privacy out of that okay Carry On go on and so you compare that with smart contract and chain signatures right so that that program can have that it can

39. 20:51

    you know do all the same things you can have a smart contract to verify even potentially you know something like some properties that in case I mean highly unlikely but if there's any vulnerabilities inside te like you can have like additional guard rails you know don't spend all the or whatever can layer yeah you can kind of can prevent the secret key from getting used outside of the rust code or whatever code it is that I

40. 21:33

    originally chose to delegate my dcash to then additionally if that were break how can the smart contract protect something layer Defense work yeah so on here we can actually have like the private key that the G has doesn't need to be a full access key if can be a limited here which only allow us to for example call me a function call on a smart contract right function call can verify that hey maybe like only 10% of

41. 22:05

    money can be involved in a single transaction or per day or per hour or whatever and so this would allow to like limit you know potential uh liability there again this is in the unlikely highly unlikely event of somebody literally you know figuring out how to break the teage uh so far the only known vulnerabilities are all around like very sophisticated you know side Channel attacks uh which require a lot of like kind

42. 22:38

    of a lot of way to coordinate what actually the program does inside the T itself and and like oblivious Ram actually fixes a lot of that so like if you use oblivious RAM for for key storage you you like all of those atts effect um there's no known atts effect okay so I'm not 100% up to date on the history of attacks on tees although I was at one point I think but um but I really liked the part about the smart contract enforcing a sort of fine grained

43. 23:20

    capability this this there's this computer security Theory called object capability Access Control control that's all about that and something I really like about it is that it ideally merges computer security sort of layers of defense and all this infrastructure issues ideally the object security capability model merges that with the user's intent right but that's a question of UI like does the does the

44. 23:52

    user know how to tell the computer what they intend or what they don't intend and um I think that's a really rich area for improvement there's a whole um there's a thing called the norm Hardy prize Norm Hardy was this great computer scientist who I knew in person because in computer science we are privileged to rub elbows with the Giants upon whose shoulders We Stand um I forget Which Giant said that first but anyway so there's this there

45. 24:22

    was this great computer SST named Norm Hardy and there's a prize in his honor which is all about the specific question of secure user interaction design and I think the most important or almost the only question in there is how does the computer know what the user meant and how does the user say what they mean or how does you could go the other way like how does the user know what the computer meant but I think that's generally the wrong direction to go like for

46. 24:52

    example um I was using app. near- in.org and the Brave web browser and my zcash um wallet today and at one point one of them I guess it was the brave web browser popped up this dialogue this is not a criticism of Brave or of near intense this is just like it's an early stage and we all have a lot to learn and improve by experience so anyway braveway BR the the wallet built into Brave popped up

47. 25:24

    this like dialog box it says app. near- .org is asking you to sign the following transaction and then there was this giant blob of like Json and I actually I don't know if it was Json or what but I actually read the whole thing it's like a timestamp there's a hash there's this and that and then there was the part that's the intent part right I intend to swap this many eth or this many Z for

48. 25:47

    that many eth Etc um and I keep thinking this is something that somebody should apply for the norm Hardy prize and get a whole bunch of money fix it for solving this but I keep thinking this is entirely the wrong direction that the computer namely Brave is trying to explain to me what near app. near in.org said to Brave in so a it's very hard for me to understand it like maybe one of those hashes there was a non I was like what does that non do maybe there's some kind

49. 26:20

    of malicious mistake in here I'm going to lose all my money because I misunderstood what it was telling me um and you could imagine Brave and near trying to explain it to me in more understandable terms you could have an AI that like reads it off and explains to me what the pieces do or whatever but maybe that's just entirely the wrong direction the computer should not be telling me stuff

50. 26:45

    I should just tell the computer stuff right I told it I wanted to swap tinz for one eth or whatever it was I just want to know if you understood me correctly computer don't don't tell me about all that other stuff anyway just I'm just rambling but yeah like the computer should be understanding the human more and not trying to explain stuff to begin I mean that is where AI is going

51. 27:10

    effectively I think I think the next evolution of UI and and we actually working on some of that like we're trying to reink like okay well if you imagine you start some stretch and like you throw away everything all the UI stuff moved down so far right right now kind of all AI effect just just end up in the chat right like you're just talking to them but obviously chat is like it's not a optimal solution for every single interaction because you

52. 27:40

    want a better presented artifact or a specific thing like if I'm ordering a taxi I want to see a map to see where it is like it is map yeah yeah good point you want this kind of Rich elements but this Rich elements should be coming back from we call them agents who actually doing the work right so in the case what you're describing in a way it's like the agent in this case the solver that is effec understanding your intent they should return you back a UI that kind of

53. 28:16

    actually explains to you what you're get right I guess there is a little information I did want which is how much I would pay yeah well which again like the you actually showed you that the problem the problem where the is is like the your wallet versus the app is different things right right right like the way the way app. near- in.org explained it to Brave and then Brave wanted to explain it to me to make sure

54. 28:41

    I was okay with it yeah that but this go back to I guess I'm actually interested in intense after all now but dude I mean this goes back to you know there's a uh wallet right that you use and that wallet can embed the intent protocol itself right like near intent is just the demo UI of the protocol ideally the wallet that has the capability of you know like capturing user intents would use that protocol directly instead of having this kind of the the the website explaining the

55. 29:23

    wallet what need to be done so I think that's where one of the opportunities is yeah I and there are some user interface there's some live user interfaces that you can play with or whoever is building this can play with to learn from or or the makers of those user interfaces could upgrade to add this I'm thinking of the one that I've used the most often is Edge wallet Edge wallet it's a multi-coin wallet which includes shielded zcash and ton of other coins and it has a builtin

56. 29:58

    frontend user interface to those centralized uh instant exchanger services and it does something that's a lot like what near intense is doing uh but less but worse but different um but it's a really good to learn from because it's really a very like the users intention and desire at the start is really the same and the that all the fact that you don't leave the edge wallet is a really big user experience upgrade right you like you

57. 30:35

    never go load a different web page and cut and paste or scan anything from any other web page so you load up this one app and you just keep that one app for granded on your phone the whole way through and you tell it I want to swap 10 eth for one Z or whatever and then it does a whole bunch of clunky insecure sketchy stuff which is it talks on the back end oblivious like without bothering the user too much with

58. 31:04

    this it talks to all these instant exchangers like let's exchange doio and several other ones and it says to them how much would the slippage and the fees be if I Ed you and so forth and it picks one and then it sends my funds to that one and then hopes that that one gets back to us with the other funds later but okay so the current like you said it was the app. near- in.org is just like a demo app of the underlying infrastructure I agree as a fan of

59. 31:37

    infrastructure I think that a lot of the near ideas are really cool and interesting and promising but what Edge wallet has is that they've already run you know tens of thousands or however many hundreds of thousands of users through this flow over and over and dealt with probably a zillion complaints and objections and suggestions from all the users so uh okay so my point is I guess Edge wallet should look into near intense and vice vers yes but also so

60. 32:10

    is I use zashi as well right ideally all the wallets can can embra as well yeah zashi is actually my favorite daily driver wallet which I go to first like if I'm on boarding someone to zcash buying coffee or buying Le I I go to Chipotle and get a burrito bowl and pay from zashi um but it's only it's currently zcash only so it can't do a swap without leaving without leaving the app now I did use zashi for the source of the swap

61. 32:44

    I did today because it has my Z in the shielded pool so whatever I do with it basically it retains my privacy in the sense that whatever I do with it can't be linked back to the original source of my funds or my other that cool or anything else I did with it so here's actually the interesting thing because intense effector just need any cryptographic signature so if you start with zcash uh you can actually dep like you the deposits are just a transfer and you

62. 33:16

    can actually do all the trading in on intends with just signing a message like it's not actually a transaction you didn't need a near account to trade on intense like you can use you know and so in theory that actually can actually offer that experience where you can through your zcash wallet and zcash key hold assets on like kind of off other chains and then if you want to withdraw it to the real wallet like to the actual

63. 33:47

    Bitcoin wallet you can do that but is interesting so what you're saying let me Echo it back but kind of objective part of it or refine it um so you're saying is because of the near chain signatures thing and maybe the intents and maybe the tees and all the different Technologies all the background attack yeah that a wallet that has no secrets other than a zcash spending key can control a smart contract running on year which controls like a Bitcoin spending

64. 34:25

    key is that right that's really cool but my injection is you still need DUI like even if even if your wallet that you're looking at doesn't have any Bitcoin spinning Keys maybe it doesn't have the code to generate signatures on using Bitcoin spinning Keys that's fine but the the other part which is really important and critical is the UI like does it show you a Bitcoin balance does it have buttons that control your Bitcoin bance so it's interesting but yeah for sure so I mean

65. 34:55

    that would be the integration for what that you to do which is like adding adding sound birth displaying assets and and so for people that don't know zashi is this really great zcash wallet and it's made by the zcash organization that I currently have no affiliation with uh the electric Coin Company it's called they're a super great high-powered team which I founded and built in years ago but I'm long gone from directly working with them and I work for this completely other separate thing called shielded

66. 35:28

    Labs um so I can't speak for zashi but yeah that would be awesome if zashi added all those features and I wouldn't be surprised if they would because they've for the last year or so they've been like heavily investing in user-facing quick feature upgrades like they've added completely new things like the fact that you can buy your burritos from chapali which is my favorite feature um you can also buy all this other stuff bedb and Beyond and Al um makeup and all

67. 36:00

    those other like retail stores but um and they recently just around Christmas I think they raced to have this shipped Before Christmas as a Christmas present to the zcash uh integration with a hardware wallet so you can keep your signing key even further offline on a on a physical Hardware wallet device anyway my point is the zashi people have been what rapidly adding user facing features so I wouldn't be surprised if if the

68. 36:32

    users were demanding it if they added different coins and near intent based swaps and all kinds of stuff but I can't promise anything like that because I have nothing to do with them well that's that's a decentralized world right we just uh we just highlight the direction and then you know hope that uh other folks follow so maybe switching a little bit gears um because I mean I've personally

69. 37:07

    been doing a lot more work on AI kind of we are using the same intent architecture called the AI um interaction that's why we're building this kind of te infrastructure as well because we think that you know otherwise like doing stuff on chain and mean just overhead is massive this actually gives you a very good balance at like like verifiability and actual practical confidentiality which no other solution offers you for actual AI workloads but

70. 37:38

    one of the challenges obviously we've been discussing is that all the transactions are completely transparent and so although you're you know you have all your AI um kind of Agents running inside confidential Computing and you you don't no knows what kind of the actual actions or inputs outputs are going but if all the transactions are transparent then you know so I mean I this can be like from ordering pizza and burritos to uh to interacting you know a

71. 38:12

    trading algorithm that you know actually makes trades and like if everybody sees all the trades linked together obviously you can start front like reverse engineering algorithm and actually front line which is highly undesirable so I mean we chatted a little bit kind of Asing around some of this like how Shi pool and shield transactions can be useful uh there's some ideas we've been discussing I posted this idea because we

72. 38:42

    actually have a one trillion agent hackup Ono where kind was the idea that you know we G be a lot of agents in the future and one of the ideas was like hey can we have an agent that effectively like shuffles the money back and force between you know take and zcash and although they're you know um I mean we can discuss the exact kind of concerns with that solution it can offer some

73. 39:10

    practical you know hey you you put some money into this pool it kind of shuffles it in and you know over time kind of accumulates it back on different addresses and stuff like this um but yeah I'm curious kind of how you think about this world going going in the future or especially also tennis interaction between ten zcash and what are some of things we building on the iso well my short answer is definitely not going to work today maybe maybe not work

74. 39:52

    tomorrow like there's a there's a super common just in concept on the part of people when they think about privacy which is they as most people assume that privacy works like what we call a mixer that like they let me explain what that I mean by that they assume that you're going to start with your asset over here where your enemy can see it either because it's in a bank and your enemy has access to the bank or it's on a blockchain and it's a transparent blockchain or whatever and what most

75. 40:32

    people think they want is I want to take this and do something to it to make it private and then move on after I've done the thing which is what I tend to think of as the mixer Paradigm like the thing in the middle is a mixer that's going to add privacy or privatize or or break the link between the first step and the third step and my first answer so I will respond to your idea and refine it in a minute I mean ref like question my own cty in a minute but

76. 41:06

    first let me say 100% certain that's just never going to work like um you can't add privacy in the middle step because privacy is what you lose when information leaks and you leaked information in the first step and the third step so even if the thing in the middle is like the world's best thing for not leaking any additional information it doesn't matter because you already leaked enough information in the other two steps to lose your privacy

77. 41:40

    so even if it leaks zero additional information because it's like the perfect black hole it doesn't help so that's the first that's like the simplistic answer and and that's absolutely what's true to that like um there's a lot of data points I read all the cour documents and newspaper articles and police reports and like uh security analysis like audit reports and all kinds of stuff all the time to try to make like double double check my understanding of the security stuff

78. 42:12

    because it's like my favorite hobby is learning about security stuff and there's a lot of examples I can give where people had that had that assumption that what we need to do is break the link between starting ending point and we're going to get some kind of technology to do that for us in the middle and they 100% failed here's a my favorite example there's this instant exchanger named shapeshift or there was I you went out

79. 42:43

    of business but or the government put it out of business but for quite a few years maybe mostly before you started you said you're in year four of your crypto battle since May net four years since May net oh four years since main net okay yeah so that makes you a battle scarred veteran too but so you may or may not remember that for a many years in the early days of crypto shapeshift was the dominant instant exchanger was by far the most widely used instant exchanger

80. 43:14

    and instant exchanging was one of the most widely used use cases in the whole world of crypto at that time and um some computer scientists security researchers from University I think it was University Cambridge London which is a really good uh computer science research University they looked at the logs of all of the users ever of shapeshift for like three years or whatever a lot of logs right and they said huh the most common thing anyone ever did with shapeshift for the first

81. 43:46

    three years was they brought some Bitcoin or Litecoin or whatever they swapped it for Monero or zcash and then they swapped it back to the original thing they brought Bitcoin swapped it swapped it and get left with Bitcoin so a they paid some fees to shape shift B they were probably trying to gain privacy and see they gained no privacy because our grad students Just DE anonymized all of them in order to get a paper publish um so my point is until now that

82. 44:19

    has never worked and the current state-ofthe-art of that is Monero not to not to be negative but Monero basically has a similar architecture where there's like a step in the middle maybe that maybe it's different but my point is everything that any and tornado cach that's a very popular one very widely used very widely understood and it never worked like no almost nobody got almost any actual privacy out of tornado cach it's a PhD researchers from UCL can

83. 44:53

    probably de anonymize all the 90% or 99% or something of all the tornado cach users just to get a o why just for the very simple reason that the information emitted at the beginning and the information emitted at the end is already too much information to keep you safe so it doesn't matter how good tornado cash was about not emitting additional information in the middle okay so if I stop there this like takes us up to 2025 that's I think that's the current state of the art and but there

84. 45:27

    is still hope in 2025 which is privacy doesn't come from money in motion privacy comes from shielded money at rest and we have the example to prove that one too which is the zcash holders who have zec in their shielded wallets if they swap it through let's exchange doio or much better through app near- and.org for ethereum or whatever there's zero information like the first step leaked zero

85. 46:00

    information and then the last step which is now you have eth in your ether wallet that leaks a whole bunch of information about how much if you have when you got it how you paid the fees or whatever right but that's okay because there's zero information about you or your original Holdings or anything else you've previously done so curent St of the art privacy comes from money at rest

86. 46:24

    and just you cannot add privacy to money in flight but the question is will that change in 2026 because of AI or something I I don't know maybe I'm I think it reflects on a biger question so let me talk about the narrow computer security and privacy thing that I have over specialized for 30 years or however long I've been doing this very narrowly uh definitely doesn't work today in the future if both like the adversary has an

87. 46:56

    AI that's trying to analyze you like today's analysis and Arkham and whoever else and you also have an AI That's trying to protect you which one will win I I really don't know I wouldn't bet too hard on either one because I think it's easier for any kind of intelligence human or AI to detect patterns than to generate fake P enough convincing enough fake patterns to confuse the other guys I think I would bet on the Spy still at this point

88. 47:38

    but maybe I don't know that's complicated but but actually okay I want to let you respond but then I want to zoom out to something that you've said that's really inspiring to me okay what were you gonna say yeah yeah I mean we we should talk about the higher level but I mean so I I agree with you that like if you literally put you know $5.7 worth of Bitcoin in into zcash and then withdraw back 5 point whatever 69 you pay the Fe

89. 48:03

    like one C the fees yeah that is highly traceable right now IM imagine there is a pool of um let let's just for Simplicity like dollar dollars right and it exchanges them into uh a fixed amount of zcash right you know like it's it's it's a like a cash Bill s z cash right like you know one like on into on and it's and it's doing that like constantly just like at all times there's always like uh that

90. 48:40

    process happening now I have the user can come in I can deposit more money into that pool and the registry that I deposited is Thor t right so the registry itself is encrypted and stored in Te and this pool is just continuous like moving things around maybe but the first as an attacker you know how we in cryptography uh like to think like the bad guy so as as the the attacker in this

91. 49:11

    scenario my first move is just not even to look at the thing in the middle right like it's it's got encryptions your knowledge proofs tees whatever I don't really care about that I'm just going to look at all the behavior that I can see of the users that start the behavior to come you know over time but there's a lot more that I'll do as an attacker and this is something we've learned from cryptography is it

92. 49:35

    attacker gone no no I just wanted to let's finish the the the the structure so the structure is there's a pool that withdraw in depositing but it uh when it withdraws right outside it puts it always a new address so you always have like one zcash worth of dollars on new addresses on there and so when I want to spend and some like it's effectively instead of money and rest in zcash I'm trying to keep money rest in dollars but they are effectively been

93. 50:08

    split like they effectively been mixed through the cash and split into individual addresses and so only when I need it I kind of joined it back like only subset that I need to spend I joined back from and then spend it so in the way I'm I mean I'm trying to use with cash kind of ZK part and then t as a like record detailing of the affecting the shielding uh right and and well so far it's kind of sounds a lot like the

94. 50:41

    tornado cach model oh by the way I don't know if you know but um just before they got shut down by the different arms of the US government they were in the process of launching an upgrade to tornado cache which allowed the users to uh keep and transfer for their assets without leaving the shielding pool which would have been a major Improvement which which yeah yeah which is the main unlock effectively yeah like if you can because now like you can yeah keep

95. 51:11

    everything in your cool right right um okay so basically I'm pretty skeptical like the thing you just said no offense I think you're super smart and you know know a lot of things and have to come up with a lot of good ideas uh but the one thing that I'm an expert at is this one little thing and so far your idea doesn't sound very promising if I had to go put a bet on PO market today I would say oh yeah the chain analysis is going to rip that to

96. 51:43

    shreds all right well we need to keep thinking about itbe I think there's yeah oh but something you said really inspired me I I think there's a really awesome like I was thinking the other day because what you said was something like what's the user how does the user interact with an AI you and I or at least I and probably you are assuming that there exists an AI whose only interest is serving the user like

97. 52:14

    there's a whole different dystopian world where the AI works for Google and I think it works for me but it doesn't actually serve my it makes money for manipulating me Google that's let's assume we don't go there I'm definitely not going there but um assuming there's Ani that works for you then I think it's really interesting like um I was I was thinking about this progression of like web 1.0 to Web 2.0 where what am I think I don't know if this is accurate or the right

98. 52:45

    terminology but like what was I doing I wanted to buy something or find a store that was open or something the other day and I was thinking the way I'm going to go about this is ask the brave browser search engine for information about the store and I'm going to navigate to the store and then I'm going to have to try to figure out how the store's user interface Works to find if they have the

99. 53:09

    product that I want oh it was light bulbs I was looking for light bulbs yesterday it was a pain in the butt like first I went to Amazon so I didn't even go to the search engine I went to Amazon I was like show me the right kind of light bulbs and it gave me all the wrong results like the wrong kind of light bulbs and I tried and tried and tried I they even have a built-in Ai and on I was like anyway that didn't work so then I went to like a brave browser search engine or Google search engine something like can you show me where to find the

100. 53:33

     right kind of light puls that didn't work and then I was thinking you know this is going to be really really radical because this is ineffective and a lot of people have a lot less sophistication and patience than I do for dealing with this stuff like a lot of people can't remember or tell which website they're on right I don't know if you like I'm I'm giving Tech Support to like my mom this classic computer geek

101. 53:59

     example of the computer geeks mom but I I give people TCH support often and they don't know which website they're on really they don't they don't know the name of it or the URL or what website it is they're just trying to get their question answered and they're just waiting for it to see an answer um anyway so but this is really promising if we get a user loyal AI then it can give them what they actually want which is just find me the damn light bulbs of

102. 54:30

     the right kind and I don't want to know which whether it's Amazon or bulbs.com whether you use Google search or Brave search or anything else about it I just want the light bulbs and that really matches what you've been talking about about intense which I still don't really understand but I'm starting to get slightly interested and um and it'd really be revolutionary and it happened

103. 54:53

     because currently the uh the network effects that allow big tech companies to monopolize things I think are completely blown away by my mom having a trusted agent that she can ask to get her whatever she wants really eliminates or completely changes all of the network effects it's really interesting anyway so maybe I think AIS are not really going to help with fixing the mixer I don't think the mixer will

104. 55:24

     ever get fixed even by AI but I could be wrong but I think they might help by fixing like the Google Facebook Apple Microsoft Amazon yeah no that's exactly what they're working on um and uh we do we do need also privacy on the user balances and and kind of spending side right no I agree it's a no I I agree it's a really strong it's really important to um pay attention to what the users actually

105. 55:50

     want instead of telling them they shouldn't want that so there's a lot of users what they want is to take their Bitcoins privatize them and still have their Bitcoins I just don't think it's possible but it should be on us to figure out some thing that we could either offer them a product or teach them or something um but in the meantime the the early adopters who know what's what are just holding Z

106. 56:21

     instead yeah uh but yeah I mean I totally that that's been kind the main goal is like how do we how do we enable like was this platform check that's happening from indeed using Google and Amazon and effectively all those products that are there to make money on you to what your assistant will look like right and actually for your assistant to be really uh useful it needs to know as much as possible about

107. 56:53

     you and so you kind of naturally this becomes the most knowledgeable thing that has access to all your information to all your banking crypto Etc and that is I mean you know makes everybody uncomfortable when you describe it that way and so privacy is extremely important in EP GES who cares absolutely yeah and I think we can learn a lot from modeling on human relationships especially

108. 57:24

     lawyers doctors um executive assistant so I have been privileged in the last eight years since zcash launch to have lawyers work for me and an executive assistant or two that have helped me remember my meetings and answer my emails and stuff um and I really I think most people haven't had that experience right Mo most people have never had a lawyer work for them um but I think those who have

109. 57:55

     should pay attention to like the norms and the relationship because I think that's probably the best indicator of what the relationship between a person and their their AI will or should be like um and the number one rule as far as I've noticed working having worked with many lawyers in many contexts for the last many years the number one rule is what they call the duty of loyalty is

110. 58:20

     this really important Norm in American at least legal culture which is if if a lawyer works for you it is 100% unacceptable for them to also work for anyone else that might interact with you right like if if a lawyer is your lawyer and they have some kind of side deal with the with the other person that they didn't mention to you that is so unacceptable that if anyone found out

111. 58:46

     they would just lose their license and never be allowed to work as lawyer again might go to jail depending on the details and that's that's a really really strong cultural norm in American law and I really think this was inspired by a um venture capitalist named Albert Winger many years ago who was quite ahead of his time I think but he wrote a blog post or something that I read which

112. 59:10

     said that needs to be the norm for AI like if an AI says I'm here to help you tell me information about yourself your intentions your private information whatever and I'll help you if that AI has any loyalty or in or side deal with Amazon or anybody else that might be benefiting off of you that's completely unacceptable that that should never be allowed yeah so uh we need to we need to work toward that the that's not the current

113. 59:44

     state yeah yeah scary even even the model training process itself like even with open source model they actually not open source they open parameters you don't actually know one so we're approaching an hour so I just wanted to kind of round up I mean there's a lot to continue kind of innovating on both sides um we do have in our AI haathon we do have a zash aiy for people doesn't need to be what

114. 1:00:15

     we were just discussing is it going to try to fix the I mean you can do other things we can learn from the attempt it can it can also just you know be the one that uses your zcash address and spends it on things that you want to spend on there's there's many different different things to do and then obviously the other thing is for un shielded zcash you can also now do a lot more on with it on near because actually

115. 1:00:45

     as you deposit it through intents it's actually becomes a token on in near ecosystem so you can do borrowing and other defi stuff as well so I may I'll try that next the main thing I wanted to do a swap and I've successfully done that but I've never before I've never bought an NFD yet and I've never lent or borrowed any crypto yet um maybe I'll try that next yeah sound good well bringing zuka more on chain that's uh I t-shirt for that I've been waiting I've been waiting until I could do so

116. 1:01:19

     without linking it to my like through name and home address and whatever sounds good well it was a great conversation thank you very much excited excited to continue working together yeah likewise all right looking forward to next time