DevHub Live 29 - Building for Privacy with Calimero & Frol

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

1. 0:26

   what you want you want you want what you what you want

2. 1:28

   want me want want want hey hey

3. 3:13

   you hello and we're live good to see everyone again it's been uh hectic with all these events redacted Dev K and eth Global but we're finally back um for another episode of Dev Hub live so I'll bring the uh the guests on now we have good to see you guys Sandy from calero and everyone know fr the infamous FR from from Dev Hub how are you doing guys amazing amazing thanks for having us hey Sandy long time to see you long time no

4. 4:10

   see I just to like fix my haircut but I'm good I'm good H looks it looks perfect better better than mine not not the backwards cap uh but yeah great to great to meet you Sandy I don't think we've uh we've met before but yeah it'd be great to get into privacy I know that's your speciality over at calero and I know that you had some discussion points for all that you wanted to get into so I'll let you lead

5. 4:44

   the charge um maybe you can give us a bit of introduction should I start yeah sendy go ahead okay so I mean like long story short I'm fraud used to be my boss my manager so we have a long history together uh I used to work at near uh I joined the infrastructure team spent two years there and then like calimero started as a spin-off with the idea to bring private data onto blockchains

6. 5:16

   um and I mean initially we started this kind of private Charing approach which was like let's build side chains which um like enhance near uh in a way that like you can run your own side chain which like all the transactions are between the peers there uh and then we try to build some applications on top uh which we thought are cool like you know like building chat protocols maybe rep basements for J and other things uh but like the problem which we faced in

7. 5:47

   general like the blockchains are built for mostly transactional data um and if you want to really build something real time um you have a very hard time believe me uh so we start with this kind of second idea of like caliro V2 uh which was like okay let's improve the consensus uh and like keep the peer-to-peer interactions between the people still um and this is like how like calimero what it is that this form is so it's essentially like a very fast

8. 6:16

   crdt with like confli resolution between parties all the data is encrypted all the data is stored client side so like if you use WhatsApp in like normal world like all the data on WhatsApp is sent to us centralized server but then stored on your device and like calimero is dead without a centralized server um so like this is yeah long story short yeah I want to add that uh s started under selling uh story of the cero but I must

9. 6:49

   say that it was a very ambitious project from the go and there are a lot of challenges on the way and I hope we will have a pretty full discussion today regarding those challenges just uh very open and uh direct I guess uh I hope we can that would be great content for many uh who are listening in today to learn uh what we have in mind and how the how the challenges are what their state and um

10. 7:20

    are there any solutions for those problems right s yeah I think um the the Privacy topic is really interesting to me and I think it's something that in blockchain we all are here for and we all align with but it's not something that's so ubiquitous so I think finding Solutions and and ways to bring privacy like on chain very interesting to me and hopefully to a lot of others here um I also find it amusing to hear people AB

11. 7:56

    accusing blockchain in U like uh cyber criminal space uh being used heavily while at the same time all the transactions all the history is crystal clear and recorded So it's the the most opposite of what can be out there like in centralized uh Banks you have no that visibility uh by default uh only like maybe some actors can have access to it but in blockchain anyone can actually

12. 8:27

    review anyone can build the uh relationship and and and um um you know uh chain of uh events uh and back trace it at any point in time uh but at the same time I think uh there are use cases where we want to preserve uh some privacy and uh um I want to say that uh the on the surface uh there are several um problems and S and possible solutions that I know uh for privacy and that is uh first we we want to preserve the data itself that has

13. 9:15

    been um uh sent over uh that um even though it's open in like if if you post it in the blockchain it's open and anyone can read but if we if we encrypt it then whoever encrypts whoever reads it doesn't have clue what's actually inside of it but there is another problem of uh even fact of um transfer fact of communication um that is a different uh problem like when my account uh

14. 9:47

    transacts with the chain that could also be an indicator uh and leak my privacy um Sandy What's um um what's your take on it and what what kind of problem to uh solve I mean so like elas said it like a long time ago like blockchains are built to be transparency engines right uh and like once you started designing back then like you didn't have that much of privacy enhancing technology when like five six years ago like most of it was

15. 10:23

    like very theoretical it was there but slow um and like now like I mean you build a blockchain to be a transparency engine for everybody to validate the state right but like now we see different types of Technologies to be able to enhance this data I kind of like separate it into different categories uh first category is like where do you store the data right like that's the question um like where is this data

16. 10:48

    stored and like that's the main kind of the problem of which we are trying to St like Sol with Caro like no if you're storing it in a centralized server you know yeah it can be encrypted but still you you have like censorship risk like somebody nukes your data you nothing you can do about like somebody pulls you from the platform nothing you can do about it uh then second part of it is like uh I mean interacting onchain

17. 11:12

    with privacy right uh so like you obviously we have zero knowledge technology these days which has to use some data to provide proofs of of of of of of certain things I mean we had zcash and Monero for transacting right so like if you want to have payments I think this problem is kind of solved like you can transact privately uh but then the second problem is like if you use zcash and mon eror you cannot really use

18. 11:37

    exchanges you cannot really uh you know like offr regulatory questions uh arise so like that's one thing and then the third part is like private computation part which we kind of try to work on as well so like you have many projects popping out nearest using tees for like trusted execution environments to build you know private data model models uh some people are trying to use like you know fgs like fully homomorphic

19. 12:05

    encryption which in my view is like a very very good way to do private computation the only problem is like super super expensive computationally wise um so it's like not feasible for every single thing um but there will be use cases for that and like same with NPC right like you can build decentralized networks which use NPC to have encrypted data compute so like these are like main three categories right like how do we interact and store

20. 12:30

    the data how do we interact with the blockchain that we you know preserve privacy there as well um and then like how do we do like more heavy compute on things like Spa filtering on private data uh AI on private data um um and other things um you can touch on every single one of like what are the shortfalls what are the hard problems there but I think that this kind of majorly covers the Technologies we used today except like you you know the old stuff for like encrypting traffic over

21. 13:03

    the wire I think we we kind of solved that a while ago so and how like when you were starting calero how did you navigate the different solutions and then come to like work on the one that you're working on now like why did you decide Your solution was the best fit for what you were doing um I mean so initially like the story is uh I I I I I really wanted to do something with like peer-to-peer

22. 13:34

    networks and private data and I came to Ilia and I pitched my ideas to him and he didn't like any of them um and they had this private chart project in the Shelf uh and I was like then I'm going to pick this and like like work on this um and I started working on it internally in near with like one of another person in near and then like there was a proposition from me on the table to spin it off as a separate company and like to be more efficient um and then like the thing is like once we started working on

23. 14:07

    it like I mean I was really I'm not really a def5 person to be honest and I think like um like I understand it I think like there's a value to stable coins to private payments and all of this stuff but like I always felt like when I joined like near and I joined the web3 space um that was like one part of the reason like why joined I I used to work at Facebook and Google and I have a big this this hate not I have a big hate for these companies um now that I've

24. 14:39

    been there because like I there's been endless times when data was compromised from the users uh like they preach one thing and then they do other things um and I was more interested like how do we improve you know like other parts of like social interactions between people chat communication you know J replacements and like all these normal applications which are not maybe related to

25. 15:03

    defi um and like we tried to build like few prototypes and like we iterated on the architecture until we figured out like this would be the best way to solve the problem of like storing data client side because then the user owns the data make the interactions between the peers you know directly peer-to peer encrypted so you never depend on a centralized entity uh and then also I have options

26. 15:25

    to back up the data right like if the user loses um the data or the device or something like they should still be able to recover it somehow um and then like obviously the blockchains as a very good means of of of payments like for making payments and like they're also really good for like strong consistency stuff so like calimero uses blockchains for example uses near for storing member states

27. 15:51

    which have to be like publicly globally available and they have to be strongly consistent so like all the stuff which is has to be strongly consistent with kind of delay to the blockchain or let's say like you want to build a private out like that's an example which we build internally few days ago I posted on Twitter like if I want to have a three of us in a DA right and like we have a proposal which comes to us let's say

28. 16:15

    let's find this project and give them 10 year maybe we don't want this proposal to be public and our discussions about this proposal to be public uh so like this would live on Halo like we would have the discussion we would have the vote and then once we cast the vote the vote would be sent to near and then if the vote is a yes then would it would M either make a function call to a contract or it would make a transfer right um and like in the future like currently this the the payments are

29. 16:46

    public right but like if you had a project on here who has zero knowledge payments then we could call that contract instead and make a private transfer like end to endend the whole process is is you know encrypted and like nobody see what's happen happening um so that's kind of like what I'm more interested in like there's also payment stuff we do obviously with with the private proposal but it's not only the the whole component of of of what I

30. 17:10

    believe is the future right actually that reminded me of uh C chat uh I mean that's the one I don't want to talk about anymore really like I I I would right yeah it was last near con and but I still think it's valuable lesson learn that we want to really share with the community what what went uh good what went wrong and what's ugly I mean you can share if you major problem I will tell you like I mean we use the calimero chat internally as a company because like I pre like I really believe that you should do what you

31. 17:49

    preach so like if you preach privacy you should really try to get to a point where you start using privacy products right or if your company uh in web three you should probably pay people in stable Co on on chain right like if you don't use your own product nobody else gives the right or if you're Crea so I can be like you know say this word nothing for or if you're creating uh decentralized front

32. 18:13

    ends have your your like Dow Dev system on there as well exactly so like if you're a dow right and like you should strive to use decentralized parts of the stack everywhere uh but like the major problem we had and like we knew it was like user experience right if your user experience is terrible and like you like we tried right to fix it and we really invested time in like tweaking it and

33. 18:36

    making and the goal was really to make a user experience we were happy with and we got to a point where we were like okay it's usable uh but it wasn't really great but from like all of this process of like building the chat trying to fix the user experience try to really analyze like what were the shortcomings you know of like building it on on on a private blockchain um like we came up with the new system which we call V2 today which kind of mitigates

34. 19:07

    all the problems and now you can really build usable apps today with what we got to but like it took us a long time right like it's very hard to test all of these ideas it's very hard to like build these ideas right like you are testing building rebuilding um but now we're in a good state is it just a tooling problem or or there were still uh conceptual like technical challenges that were hard to Sol there's two

35. 19:38

    technical problems like there's like two there's a business problem and a technical problem first problem which we had with like the first like there also like the new this also went into thinking about the V2 design running the private chart itself was very expensive because we used the near validator stack and like we used at like modified near core version right but it was very expensive to run like you have to run multiple validators cost you like few hundred bucks a month then you have to

36. 20:08

    run an indexer if you want to index the data cost you a lot if you really want to optimize the system like you have to put a lot of money into it which works in decentralized scale because in decentralized scale you have inflation which covers the cost the fees right this gets really distributed to like there's an incentive for people to do it but in our case there was like no incentive for people to pay a few hundred bucks to get privacy right or

37. 20:31

    like use it it was the same for developers either we Pro we provide a free service for them to test it out but then for every single chart we run we have to pay you know like thousands thousands and thousands of do dollars a year or we charge them and then nobody tries it yeah like nobody tries it you know like you have to subsidize it but we cannot subsidize it so like how do we reduce the

38. 20:54

    cost I think it's like a very very common infrastructure problem I think we've discussed about like rpcs similar situation here um yes but like rpcs I can also touch on that like rpcs at this point like you have a business already right like if you already have project like n was subsidizing the RPC for a long time like many layer ones do uh but at the point where you have already like you know some solid projects then they

39. 21:26

    can start you can start charging them uh but like you you have a sun cost for like years until you get to that point um but you should still provide like you know free RPC for like few requests per second or whatever to like incentivize new people coming in it's like premium model I mean it's been around for a long time um and like the technical part which I I didn't touch on like why it didn't work like block times were very

40. 21:51

    hard um like if you have a block you have to wait for things to finalize every time and like the user experience gets a bit weird in like real- time things um like if you if we are chatting like the delay is a bit funky um and also you have an amount of things you can put in a block right so like if you cannot put everything in one block then you have to wait and then if you really have congestion uh you have to create a

41. 22:20

    new Shard for a new channel um and then the cost goes grow growing again um and it really works in a decentralizing manner because like somebody's paying for this right like somebody people are paying gas fees but in our model there was no gas fees there was nothing to subsidize this um so we went for like a more lightweight model like something you could run on your device it you know

42. 22:43

    talks to the people you want to talk it signs the messages you want to sign to like certain things right uh so like that is generally two reasons why we decided to rebuild it mhm interesting and I um I spoke to a few people from calero at redacted and one thing I wasn't really sure about is like you're saying that the user data or data in general is stored locally right um but then how do you have consensus on

43. 23:22

    things if it's all stored locally so for example I know you don't like the the defi example but let's talk in terms of tokens if I'm storing that I have 10 10 whatever tokens locally then how can we confirm that that is the actual amount that I have maybe I misund some amazing question so if you have tokens you put them on near near provides you Global consensus Calo does not do it but if you want to

44. 23:53

    store all the history uh about let's say uh your pay payments or Shopify experience like that maybe you would put on calimero like like the priv is a good example again because like the proposal is like three of us it gets replicated between three of us if we vote it gets replicated between only three of us right but like once the final decision is made this gets finalized on chain and

45. 24:17

    then the money is onchain because blockchains are really good at providing Global consistencies we use crdts they're not crdts are really good at like Fast replication of data and resolving conflicts but I locally verify the things I'm interested in like so if you send me a message I can verify that you send me a message uh but somebody else outside of you know the application

46. 24:41

    cannot do it um and like it's it's it's like a tradeoff right like what type of application You're Building what type of of of things you need right like what what it's the same in if you think about like web to right sometimes you use mongodb sometimes you use SQL sometimes you use both of them for a different use case it's the same here right like if you really need private data which is shared between peers with

47. 25:05

    like permissions and stuff and you don't care that globally uh the state is uh like accessible that people other people can verified except you then you're you're fine right okay telegram right like I just want to say make sure that if FR sends me a message I can verify that FR actually send me the message and that message was not not tempered with um and like I can do that he signs it with his key I get the request I verify was done that's it right so that doesn't

48. 25:38

    mean that cero is kind of well integrated with near Main net and that the like the point of integration of calero attack was near or Calo is totally independent I mean calimero itself is totally independent in a way that like you can just spin up a peer-to-peer Network and like you're fine but like if you want to have strong consistency then you can use like blockchains right like and we provide different blockchains Integrations like is one of our major because we are believers in near uh but like if you're

49. 26:14

    a developer building on near and you want to H private data you can um so like let's say we have Sputnik Dow currently all the proposals are public I would really love to get calimero integrated into Sputnik um so you can have like private conversations and discussions and then just votes happen like transfers happen on chain so like that's kind of the premise right if you don't need blockchain and like if you're building a like you know peer-to-peer messaging system you maybe don't need

50. 26:43

    blockchain like you can just peerto peer we interact right um and that that's it um like you three three pillows of uh like privacy that we we need to care of is cimo covering all of them is it like no no no no calimero mostly covers like uh data store and a part of compute in a way at the moment but the idea is that we cover all three of them uh and to explain how we plan to cover like cover all of three of them so like obviously we don't want to build every single zero knowledge thing we don't want to build

51. 27:22

    you know T trusted compute like you want to focus on like our core but like what we are building now in like q1 which is going to launch is going to be this like what we call Specialized nodes uh and specialized noes will be like this kind of privacy marketplace where developers can plug in other people's product and pay for for the usage so let's say somebody builds a zero knowledge

52. 27:44

    Pro for Bank history right um they can deploy a Noe which does the verification for a user and like the user just pays on near some amount of tokens the approver executes it gets some fee out of it we get some fee out of it or let's say like you build a social network it runs on calimero and you realize like somebody is posting stuff which we don't want there uh so I can say take like a Spam filtering model on near uh do inference on near and like you have a

53. 28:15

    specialized node which talks to the inference thing sends the data gets the result and say this is PM or not right um and then you like client side discard things or like somebody in the network level discards these things uh so like we want to be like a privacy network of different solutions right like we don't want to build AI right like we don't know much about AI but like near does a lot about Ai and they're like building privacy preserving AI so we can just plug it in like yeah here's users uh pay

54. 28:47

    some near tokens and you get like a monthly subscription to I don't know like spam filtering for your stuff messages or like whatever and how many notes do you Envision Cal would need so it's it should be decentralized right like the this whole network yeah so like I mean the app like when we say a node a calero node right now is like some software which runs on your mobile phone desktop server wherever and like this could be like in theory every device in the world

55. 29:20

    could have a calde running and that that's the idea but like there's first type of specialized noes like the not which are decentralized and other people running is going to be storage nodes so let's say like three of us are in a group uh you change some State Owen changes some state but I was offline in the meantime you two guys go offline and I cannot reconcile the state so we're going to have like some specialized

56. 29:44

    nodes which temporar store the state like we going to have like people who are incentivized to run the storage nodes uh to Temporary store the encrypted data until I come back so I can come back pull the data apply to my State you know decrypted and then pay somebody for doing this while I was off right like that's going to be like the first type of noes we have so we get like global data like global data

57. 30:08

    availability for calero nodes um and the Privacy is ensured through the fact that we even though we put the data in the public network uh it's all encrypted uh directed to only the people who has the keys to decrypt it and uh the data will be removed as soon as the peer is receiving it exactly exactly and like at the moment we kind of have the like we derive a shared key between members so like we have like let's say

58. 30:40

    me and you interact once we interact we have a derived shared key when we talk and like you talk with my shared key similar to like how signal does it uh and like you store this shared key for a while uh like for some time right um and that's like good enough because nobody can decrypt this uh but like we have this like we wrote this paper about like I mean we used Concepts from signal and from other people before so I don't say

59. 31:06

    like we invented it we didn't invent it like it's concept for a very long time but like we we we plan to use like Dy Hilman with like some double ratchet so we also have like we iterate the key on every single interaction currently like all the data gets encrypted with the same key in the future like you going to encrypt every single state transition uh with a separate key so like even if somebody compromises it they will not be

60. 31:29

    able to uh decrypt the whole thing right but that's a bit harder to build and not take us probably like q1 Q2 to get there as well and I think a little discussion we had before on the topic of decrypting things in the future Quantum right like in the future everything will all encryption schemes now will become well most of them will become useless um when like Quantum encryption decryption comes about have how much are you thinking about this at calero or are you thinking this is like

61. 32:12

    too far in the future to to worry about I read like somebody on Twitter posted like the estimate when like current encryption scheme will be broken it's like few years from now I don't know how optimistic is it 30 or something yeah I think I I think I saw the same one it was like when some like standard or the U will like no longer warrant it safe I mean Quantum Computing like I don't know like it's it's like you know a hard

62. 32:43

    problem to predict like obviously we already have encryption mechanisms to which are quantum uh resistable right but the only problem is they are more expensive to use computationally wise uh but usually like Tech is growing you get like you're getting you know stronger and stronger mobile phones stronger and stronger machines like more law is here and like Quantum Computing is not as

63. 33:09

    fast at least now like maybe we make a breakthrough but like if we make a break through like we are going back to Stone Age right like everything gets decrypted and um like it's hard to predict like everybody most people would say we should be fine um but like it's hard to know I mean if yes everybody moves to like flat this cryptography or like whatever thing is the best thing to be Quantum resistable and our internet gets a bit

64. 33:40

    slower for a while like yeah hard to say or maybe everybody gets a quantum computer like in in their building like somewhere under but that's that's a very valid point um that Owen mentioned because it is um um it becomes a concern uh given that there are already parties that do store all the interactions all the traffic we currently have on the internet yes in the in the hope that uh at by by the time when we will be able

65. 34:19

    to decrypt this data will be still relevant and they will be able to uh completely analyze all the data it isano but I must say we need to consider that uh me having the cyber security background I always consider that uh and when I said that you can put the uh encrypted data in the blockchain um that's half of the problem uh because like yes now it's uh protected but you you never know you can capture it now can go to some archival

66. 34:55

    node what would you you you see that it's like a bit of a problem even like now right um and not just protecting for the future so like what what are your ideas what would you do here Vlad well uh on the internet there there is this rule when I grew up uh the the rule was you do not put anything on the internet that you don't want to be there like in the first place you you prevent

67. 35:26

    prevent as as soon as you put something out there assume that you cannot remove it like anywhere and your search history as well you never search something you would not put uh yep like that's a question right like all your traffic goes through the ISP if the ISP stores all your traffic and they decrypt it eventually your F uh if Google stores your data I mean Google knows your data already but even if you

68. 35:57

    use VPN then the VPN provider uh can store the data um like even over the wire people can snoop on your data some like router which roots your traffic like even even worse I remember the times when the uh Wi-Fi encryption was not very popular it was a penalty on the performance that's why there were a lot of open networks uh in the wild and uh basically anyone sitting in the cafe shop next to you could just open up the

69. 36:35

    sniffer and see all the traffic and hgps was not that popular back then uh or available to everyone and that was a hazard a security Hazard and still we survived so it's um not many people on the internet back then right like now governments use it uh police uses it your firefighters use it you use it on a daily base all your bank accounts like you know uh you used to pay like back then you paid in

70. 37:09

    cash you didn't talk over phone with your friends so much you meet them in person but also yeah exchange right fair but also we need to be realistic and uh um the even in cryptography in cyber security uh field the uh cost of protection of the data should not exceed the cost the the price of the data itself I agree I agree with that I fully agree with that I I agree with that that that that's why we we need to have like the the there is a need for I know

71. 37:47

    maximum security paranoid mode uh for some aspects uh but I think uh we could get away with just going a little bit back to uh to the world where we use nicknames and just unique identifiers of the user instead of real names on the internet and that would already solve a lot of problems uh I mean fingerprinting right like well I mean isps used to do like deep packet and they used to do they still do like governments do like

72. 38:19

    deep packet inspection even on en crypto data right um and like the question is like I mean not like it's it's a very hard like it's also the question is like what does the new future generation look like right um that's a very hard question like I mean many people tell me like ah new kids don't not care about privacy at all they put everything on Instagram and Tik Tok and like nobody will care about privacy in 20 years uh

73. 38:48

    that's a valuable point like I don't know what happens but maybe people become extra paranoid because like AI knows everything about you um and like I mean it's hard to predict like these types of thing like philosophical and hard to predict but like think is encryption enthusiasts and people who give you know some crap about privacy should do the best we can to build the tools to preserve it in a way um because like I mean you know like the

74. 39:22

    like where do you draw the line right like that that's that's a question um is it's like the the the topic about like oh there's terrorists out in the world so everything should be in plain text to prevent terrorists right uh and I agree we should prevent terrorists but drug dealers right like like there is there's there's a line but then like should everybody all suffer uh because of you

75. 39:46

    know like I would not be happy that everybody sees every single thing I searched on Google uh like I searched some things which are like we don't want to know I think everybody did at some point like keep that private we don't want to know yeah but I think like majority of people if you ask them like oh do you want to give your whole search history ever in life or maybe not maybe

76. 40:10

    not right yeah there used to be a world with a lot of offline apps um and nowadays it's not anymore um I think we gonna go back to that to be honest um um I I just watched like the the guy from a6c uh Chris Dixon person who's like the general manager at Andre wrote the book um read right on and he says like the internet in the 90s was actually much better because people like the ideas we we kind of try to do now like they come from you know 70s 80s

77. 40:48

    now like The Cypher Nomicon ideas like the internet is decentralized like all of these ideas people thought about but then for reasons I don't want discussed probably it's cheaper to run centralized Services um and it's more convenient and it's easier to build like we went to centralized Services uh but the internet when it started right like you know irc's were like hosted by some people

78. 41:11

    you connect to an IRC I mean like in my view email is a decentralized protocol right like somebody hosts an sppp you connect to somebody's server you pull the data right um just now this is hosted by Central entity but the protocol itself is built in a way that like I host my you host your we talk to each other that's it um but like yeah convenience is what maybe was the reason uh like we didn't

79. 41:43

    have enough computer right computer was expensive um band white is also a problem right like now we're getting to a point where you have infinite band white like fiber is happening everywhere starlink they're just sending out satellites where you will have like internet connection everywhere in the world uh so I think that's going to also open up a bunch of things which were not

80. 42:05

    possible before right like I remember when I opened YouTube like even today in like some parts of like where my parents live they have like 10 megabits per second now like Optics are coming but if you open a YouTube video and you're like stuck um and that's not the best right but like at my home where I have like Optics Netflix movie download it instantly so like I can download chat history like this jeta like this peer-to-peer directly I mean I I have to

81. 42:35

    say that uh sometimes I really miss old dos and windows times where the interactions was really instant nowadays there is a lot of delay like we we we build this amazing technology so many layers of interaction that just makes things so much slower in terms of like user experience but yes it's it's globally connected but local experience is suffering a lot um but it's different depends where you live right like if

82. 43:05

    you're in a high fiber optic place then you're fine but yeah exactly you're fine but still like I I find it bizarre that even like text editors became uh much slower anyway blw blw yeah a lot of it and some something I've been thinking about recently um when we're talking about like encrypting data especially with blockchain I think one of the benefits we talk about with blockchain is this

83. 43:43

    sort of like open graph where you can tap into Data from whatever application without like needing to ask so basically I can go to uh like J lingo for example see how many point if it was on chain see how many points everyone has and give them a reward based upon that um but when we start to encrypt things I suppose and if we're putting out on chain that we don't have as much access to that open graph and and and data so where do you think

84. 44:22

    this like balance lives lies of like encryption privacy versus the open graph and how valuable do you see the open graph being FR do you want to answer or should I go hard to say send maybe you start I I go I can go maybe it was a crowd question I mean let's say like we live 20 years from now um and like technology will probably be much much much better uh and like we already have private Compu on private data right like we have

85. 45:06

    trusted execution environments we have we have MPC uh we have peer-to-peer networks where data is stored between users right uh so like if you want to compute a rating you can computer rating on any of these Technologies uh the only problem is it's not super feasible uh right now uh it's like very very expensive to run especially on like big big b big data sets um and like that's also like one of the reasons I mean I am a big believer in like all of these

86. 45:39

    Technologies but I think like practical privacy at this point in time is like you store your data and you interact peerto peer and like you are fine for like many many many use cases which people care about uh when interacting with other people uh but like if you want to like do big data and Analysis and like all of these other things like you will be able to do it right like I mean near does it like now with EES you

87. 46:04

    will be able to do it but like you will it will take some time where like you can do anything you do today probably it's going to take like 10 15 years or like five years I don't know uh but it's like a solvable problem like it is a solvable problem um I think one day everything's going to be encrypted um and I think like majority of centralized services will just be the relays or like

88. 46:27

    DNS types of thing find me somebody um and that's it or like back you run a backup at home for like whatever reason but I think a lot of that would be based upon that we expect we're still like Computing the same thing so yeah we could run what we were running in the past like in an instant and uh like encrypt that if we wanted but we're more and more like processing more and more data so how do we know that that this

89. 47:03

    will not like increase in the future like yeah the thing about data like I I often have this discussion like when like when does data stop getting produced in a meaningful way uh like how much in like there a qu a philosophical question can invent anything else after we invent these five six problems we currently have uh you know like in the 1 like in the 20 beginning of 20th century like I I read this article uh I think like um Plank's constant right uh plank

90. 47:45

    invented the Plank's constant invented like bunch of pant physics things and whatever and like he talk to his professor and he was a physics student and the professor told him like ah you took the right wrong career there's nothing else in physics to be invented right and he was like what the I want to Ru physics I want to be you know there's probably like some other things and he like no no you you you took the

91. 48:08

    wrong carrier Choice and then I mean flank ends up you know coming up with planks constant uh finding some like new quantum mechanics things a bunch of new physics things open right uh and now we're getting to a point where we're also like is there anything else to invent like okay we live endlessly longevity with transfer our things like neuralink to our brain we go to Mars VR

92. 48:31

    happens a thing to be thing everything's decentralized encrypted what what other buz word do we have uh people working on to these days um I don't know not know but like let's say we figure this out like what what other meaningful information can we come up with right let's say like all the research is done probably there is something but I I cannot guarantee like maybe we I don't think everything and

93. 48:57

    like like you said the plank still discovered is constant yes but like at this point like what you know like it's hard well I think I think if us three knew the what then we would be gazillion as right exactly exactly but I mean the the information flow right like what what is like new information like I don't like we have new research articles and and things but uh I think like quality information is is less and less in a way like I think we're going to get to a point where we stagnate for a while

94. 49:31

    usually you have like a stagnation period and then another boom um it's like with the internet right like we we invented a bunch of things but you waited from like web one for to web two to happen it was a while right and then from Web Two to web three it still takes a while to happen uh so you have like this flat step and then goes exponential play out again I feel like that that curve and the the stagnation

95. 50:03

    is contained within like domains though like plane planes haven't done anything in ages they I swear planes are like the same as 10 years ago but like AI has evolved so much since 10 years ago um yes it started in the 70s right yes but I mean tooks like 50 years to get to Transformers right uh neural networks like were invented in the 70s so it took us 50 years uh to get L of just you know

96. 50:39

    working on the tooling so the the tooling is a problem so back then we didn't have enough uh compute power these days we do not have enough private compute power private uh uh compute modules um maybe we go to uh these days uh back to these days and discuss what's uh what's possible what uh what people should uh uh be aware of uh what they should already start playing with um maybe if just a proposal how about we take for example uh to-do app um and

97. 51:19

    think about how to make it like I recently did uh live stream building into do up on near but obviously uh you don't want your to-do items to be fully public um I wonder if I can pick your brain tandi how we would uh what would be the components uh you envision what they will be doing uh to achieve the private to do app I mean just selfhosted on your client device backed up encrypted somewhere

98. 51:54

    else okay and uh is it okay that that's on a high level um do is offering that's I will build it like next week I'll build you a too app on amazing amazing I'll it for you um but overall is it like do you already have a dek um is it something that is coming q1 um no no it's like live you can already test it out uh but like currently you have to run a node on a device like you have to install it run it on your desktop or

99. 52:34

    server in the future like you download the note instantly inside the browser or you bundle it into inside the app of like an Android app mobile app like that's coming in the future and I mean we use wasmer like I mean I worked at near I'm I'm a big wasm fan uh so we use wasmer as well um and like we currently support rust is the only SDK which compiles to like the V like cimo wasm um app um in the future like we

100. 53:03

     want to have JavaScript as DEC like that's probably going to be a public Bounty so you're going to ask somebody from community on devhub and like some other places if anybody wants to pick it up and build the JavaScript SDK but maybe also like sdks for python uh and and other things I know you try to do the python SDK for near but you had the limits of of size like we store it

101. 53:25

     client side so we do not care so much about size so like in theory we could uh build any any any in any programming language uh at the moment and like the beauty is like I I I learned a lot while I was at near like developer experience has to be easy right like I want us to get to a point where like you like currently you just build some code right like you build a micros service in

102. 53:49

     calimo you do not care about how the peer-to-peer network is structured you do not care about like how it's encrypted like all of this stuff like kind of happens in behind and you just build a micros service and I think I was a lot of inspired on like how near did it and I think mea worked at near as well so like we we knew how it should look like and I think near has this very beautiful experience as well like and like some other like wasm wasn't chains

103. 54:14

     as well like you not all of them but like you build a binary you compile it you ship it and it just everything below happens inside the VM um and you have so many options for people who are developers right like you don't have to be stuck with one thing you just I prefer JavaScript use JavaScript you prefer rust use rust right uh somebody likes python better use Python uh but the idea is like you hide the complexities of the system for the

104. 54:44

     developer just to write the stuff they want to build that's a value prop right uh for the developers in my perspective I I definitely agree with that statement I had a talk reducted and I just called out like use whatever tool that makes you productive experiment the blockchain space is still pretty novel and uh there are a lot of uh things that we don't know how like what we need to invent uh similarly how web one transformed into

105. 55:15

     Web Two uh through PHP Pearl and other tools that started to generate HTML because previously it was just static and um similarly here web 4 and uh other projects on here really are um just bunch of experiments uh but at the same time the tooling should be solid uh that's what uh at devhub we are we care about um that's why we uh invest time into polishing the existing is the case experimenting with python as as you mentioned uh actually there is a good promising results uh with micropython

106. 55:55

     that is tarting embedded systems um U there are some like we the the the size is fine we we we we stuck with just a configuration of host functions uh that is the we also had the pre promising results with net surprisingly I would not never even imagine we would go the direction be easily easier than python to be honest it should be but at the same time like we are at Mercy of those who are implementing the the core um and recently there has been development that

107. 56:36

     allows to build the experimental version of net compiler that allows to build uh really lean and small vasm files with the garbage collect collector because previously they relied on external one anyway that's a different I have a question for you like I mean we like near has been around for a while now and like you you your your Dev devhub and like you said your primary thing is

108. 57:00

     improve developer experience like our de developer experience is like fine but like it's not where we want it to be probably in a few more years but I I want to hear like what is your view on near developer experience at this point like what do you think could be better like how do you feel good better things all the time what there are good bad and ugly ugly things all the time that uh you it's never ending Loop um I think

109. 57:30

     rust tooling is amazing and overall pretty solid we just put more and more tests and optimize Benchmark things uh and improve the like um standard standard support web 4 making more available and uh ABI story and all this stuff uh JavaScript stack just needs to be uh get get more tests and hands-on experience uh so we get started uh uh we catch things earlier uh with uh any uh paper

110. 58:07

     cuts um but python SDK for example I believe is a very strong uh um player uh given the AI world is living mostly in Python um and python allows experimentation like python allows to to quick quick iteration on on stuff um and I feel like it's much better fit than JavaScript uh to be honest for contract development um so that's the direction um cpnet is also interes in experiment there are a lot of developers out there I have some friends of mine who are kind

111. 58:46

     of interested in in the in tag but they don't like rust um it's you know um they shouldn't be your friend yeah that's not how I pick people but uh consider uh but the the the offchain tooling is uh in deficient State I would say like the wallet selector uh near still needs more love again mostly ux uh is just uh Laing in places and I think so Monday infrastructure side as well like with like getting up and started with an

112. 59:32

     index as quick as possible but there's been some promising tools like uh like just some simple apis like near blocks um I use them recently and just makes it so much quicker to get started indexing like historical data also the the problem is that it's all fragmented and still like a lot of tooling that that you need to you know carefully plan and craft together I mean web 2 is also has also became pretty hard um you need you

113. 1:00:07

     need to figure out the host you need to figure out the zero downtime problems you need to have your own stack database uh storage uh caching layer and all like and web3 needs all of that as well um maybe minus the hosting but Plus uh you know um caring about the smart contracts and indexing the data there so it's um um it is a tough place after all uh not very uh nubis friend friendly um so I think it will just naturally take

114. 1:00:49

     time for people to pick it up um I think um near still to Yan uh as a like project in uh software engineering uh terms space um but we're getting there we just recently implemented to-do list demo app it's it's bizarre how imag get developers to contribute externally like that's something which we are trying to figure out like we are a younger project in year and we would love to have external

115. 1:01:22

     contributors um working on various things we don't have time to and like I really believe there is you know value in having a very strong community so I like Curious both of you guys I mean like what what do you think like how do you get more contributors to year like what would you give advice to me or anybody else just getting into this process right I I think well we've experimented a lot around this and seems to be something very difficult

116. 1:01:53

     and something I've not really figured out yet um but I do know that a lot of the time just throwing money at problems doesn't tend to work like this like it's great to hire like hire people and pay them obviously but I think what you're talking about more is is having a community centered around that aren't necessarily like on the payroll and I think one of the things care about the most is just making it as easy and open

117. 1:02:30

     to get started as possible then you can hopefully attract as many people without like producing some like goal that they get and then they leave like money or whatever and hopefully if you improve the developer experience and developer onboarding then people will stay because they love the TCH and and contribute because they love the TCH yeah just to summarize my observation that that you you need to make things like click immediately like as soon as you can get people to understand what

118. 1:03:09

     you're offering and what what the like what the problems you can solve with the Tool uh the better um we recently redesigned uh dev. n.org web page uh if we can share the screen should be able to sh I think uh can you let me bring it up actually yeah can bring it up there we go so the most recent version is the the thing that you immediately land and understand okay this is the piece of code this is how you get started um and this is the minimum

119. 1:03:56

     like app and you can really scan through it you can switch to rust and also understand that one so this uh hopefully um brings more ease of use and understanding what's uh what's out there what you can get I'm not completely happy with this it's it's lacking the outcome so unless your developer with good imagination uh you will have hard time a bit of understanding okay this is the code these are the functions but what it actually gives me so far you would need

120. 1:04:36

     to go to uh start course and uh like scroll down and maybe find a last step and then learn how it will look like at the end we will need to improve on that um but overall like very short video that's what I'm looking for for an next duration of our our side and I I would you have like combination right of paid bounties and like just people who contribute right true damic yeah that's exactly what what we

121. 1:05:11

     do it's a combination of just yeah monetary incentives and just sense of community um there are a lot of people who at least start experimenting with with the tools and then share their projects um with the rest of the community we try at devhub we try to also compensate um those that are really promising and to to keep people um you know motivated to to go and evolve and and continue uh maintaining the project

122. 1:05:47

     because that's the uh clear value to the whole Community uh but definitely we do not uh sponsor every single experiment out there which is impossible it's like out of and I think probably the main people you want in your community as such are people who are actually like building products and they're going to give you feedback because it's beneficial to them right like they want

123. 1:06:14

     this API or SDK to be a certain way because they want it that certain way because it'll benefit them I think some people are entertained by jumping in conversation just because they love the tech and like the design of it but I suppose if it has an influence on them then that's when they're concerned with the conversation so if you had good Tech good marketing then hopefully people will come and then people will interact because they're they're building

124. 1:06:47

     there okay I mean that's something we are now getting to a point where we have to figure out um like you know how do we build a community of devs who care about privacy who want to build apps who are privacy preserving um but it's like for me it's a I mean there's obvious things you have to do but there's also like quite hard things to answer right uh like bounties yeah you give a bounty but what do you do after the Bounty right um or like

125. 1:07:19

     people want to build on your technology but they still have to live somehow and get a salary um yeah what do do with these people right like it's it's a lot of things which are like intertwined um I mean there has to be a balance realistically in even in like open source that we use it in um inspiration model um many people many Engineers are actually employed by some of the companies anyway like Red Hats AWS um um they pay for like mazilla also

126. 1:07:59

     pays uh developers so it's not something that uh is completely you know free of charge and uh being done on the on the side uh it starts on the side and it evolves usually as into something bigger um but it's it's hard to find the the B indeed yeah well I think that's a good point to to wrap up here unless uh anyone has anything to to end on or or any more questions they want to bring up yeah I was checking the chat um but

127. 1:08:39

     uh I think there was some mostly internal discussions between the participants um yeah I think a lot of it was uh like price P Price speculation yeah but yeah yeah great to uh great to have you very far I think it's been a very interesting discussion and thinking just like of the current state and in the future as well especially around privacy it's one of the things what the website was the was

128. 1:09:14

     the takeaway where should people go to find the oh yeah yeah let me let me let me yeah we uh I'll drop it in the the Twitter chat actually I want to send it how do I send it I'll send it here uh this is It's c. Network nice this our website this is our Twitter um like if you want to contact me or like you're interested in building on calimero either like reach out to devhub um either reach out devhub and

129. 1:09:45

     tell like we are interested in building on cimo can you make us intro to Sunday or just send me a DM on like Chef Sal is like Twitter I'm on Twitter I'm on telegram uh I am very approachable just tell me like yeah I have an idea and I really like privacy and I would like to give it a spin um so like that's uh pretty much it and like our all our code is open source uh I mean like core is open source and like some other stuff so like if you

130. 1:10:16

     want to see like our issues and uh how the project is progressing and read the docs like just go to a GitHub page and yeah learn bit more I'm waiting for your to do app I'll do it next week just for you put it on your to-do list too yes yes I'll do it next I'll do it just for you um but yeah and I mean thank you guys for having me here like I really appreciate when like anybody invites me

131. 1:10:45

     uh I really love to talk about this stuff like it's it's it's it's quite cool and I mean I know FR for a very long time uh it's been nice to catch up and I mean nice to meet you Owen and I hope we we you know talk more in the future absolutely and yeah that's it and I hope I said something smart at least you I think it's PL and there's always something new to in I'm not 100% sure but like uh at

132. 1:11:08

     least one good story yeah it's great thanks okay goodbye thank you see you everyone bye bye