Illia and Toly on AI, Stablecoins, US Policy, Intents, Growth and Tech Trends

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

1. 0:00

   hey everyone great to be here I have Anatoli founder of salana joining me today I'm sure we're gonna have a fun conversation and totally I'm sure everybody already knows you so we can skip introductions but uh yeah would love to start with uh all of the exciting things happening in us it's uh it's been an interesting time for the past four years um since we all launch mayet but uh I'm sure

2. 0:30

   uh we see very positively kind of all the developments but I would like to know kind of what's your take and how you see the the market evolving now well thanks for having me first of all uh it's been a while since we've done a podcast I think it's been what do we do the the um the near whiteboard yeah series I was like seven years ago six years ago yeah beginning of 2019 so uh

3. 1:01

   yeah six years ago well um recent events I guess there's definitely been a a change in the US Administration um that seems pretty exciting um so far I think like basically like the the big change that I see is that like Hester Pierce is not kind of preparing uh crypto policy for Paul Atkins who's going to take over pretty soon um and she's been a huge advocate of having some semblance of

4. 1:41

   rules um and Paul seems to be fairly crypto forward too um so the hope is that we get some set of guidances that says like if you do this you get no action letter or something like that um effectively SEC says that those things are okay um like I don't you know like I think people are somewhat overestimating the impact of this because I think the government wasn't able to kill Krypto because Krypto was legally in the right

5. 2:22

   they lost a bunch of cases so it was kind of a a dumb effort just to slow things down and increase cost for for like startups and Founders um those costs are a pain in the ass but all they are is just the cost right progress and everything else moveed forward um so I I don't know how much like how much actual impact people will feel from that um what I'm more kind of excited is that

6. 2:53

   maybe the latter stages of this where the stable coin build p es and like a whole bunch of kind of actually bigger meteor substantial changes make it through Congress and get signed you'll see people that are like businesses like Banks and whoever the Visas the fintax actually start adopting crypto across the entire stack just you know backend and consumer uh and that will be I think a pretty huge change because I think V

7. 3:29

   as a Global Network serves bajillion trillions of volume per day or like per year in in transfers and all of those are happening through this like very crufty old inner Bank you know originator Bank the depositor bank or whatever kind of network and if all of that is replaced with with stable coins I think you'll start seeing fees for consumers drop and just everything start moving faster we'll probably see you know

8. 4:00

   crazy to think about it but like a trillion dollars with the stable coins issued you know in the next five years yeah I think I mean to just comment on on the first part I I think the removal of this you know the sword that was kind of hanging in the air uh it's not it's not it's not just the legal fees from my perspective although those are substantial and and also not clear what you're paying

9. 4:24

   for before but I think the what what I'm excited about is the found uh kind of because especially for past like couple years it was actually hard for like us based Founders to really kind of Reason even about like hey how would I approach crypto kind of as part of my strategy and like we we as a space in general we need to get to a level where it's not a crypto startup it's like a you know solving a problem startup that uses crypto and so I think

10. 4:57

    that was really hard in us for past especially couple years where like hey I want to do biomed you know I need to use some blockchain for and crypto for this I want to raise money in this way I want to you know share ownership as well and like people were not able to approach that right um I'm excited about that changing and so like we have more you know normal startups that are using blockchain crypto on the backand for their you know to solve their problems um but yeah I mean on stable coins I

11. 5:28

    totally agree I think that just veloc of money that we can achieve is uh so much faster so yeah really excited um yeah I mean I think the the interesting question for sure is also given kind of the pro business stance and uh like how how this thing's going to start leaving in uh you know if if Dodge for example uses blockchain as well and like we not just move the uh not just move some of the business side but also government side and kind of

12. 6:00

    start integrating those things uh that's also to me would be really excited yeah um that's kind of like I I think I I think this the real change is that you know you have a a startup that's like midstage for them to go into crypto right you know prior to you know in the previous administration they had to eat a whole bunch of risk and cost um because of the fees and the uncertainty and now hopefully that uncertainty is gone so that founder can kind of make

13. 6:34

    that decision much much much easier if it makes for sense for their product to use crypto they can just flip the switch and use the technology um so we should see a lot more kind of experimentation on the early stage midstage companies that are trying to figure out growth and and and value capture and all that which is was going to be good for the entire space so I think one of the like long

14. 6:57

    running conversations right obviously we've had which is you know sharting or no sharting and uh your position always been that I mean as far as I understand you know you want a global like single Global Market right and you want the fastest synchronization possible and that is effectively how you achieve you know the most liquid Market the the most transparent Market which which makes sense and I mean obviously you've guys had amazing success you know and we've

15. 7:30

    seen with with Trump token we've seen a very interesting kind of you know uh effect of that I'm curious because at the same time you know the the clearly the amm structure right where you have you know one piece like one or two pieces of state that are being hit by everybody from everywhere is you know naturally creat ball neck and like I'm curious how yeah how you're thinking about that and I'll share some yeah you know like in defi you actually

16. 8:05

    don't see that many classic database hotpots it's it's worse than that uh because uh what happens is that um like the the bigger the bigger kind of thicker liquidity markets are easier to deal with but that's not what people want what they want is the the super low liquidity yeah very thin very volatile like nobody can really actually predict the price of this thing and they want it

17. 8:36

    now and and as fast as possible and they don't want to retry and they will turn off every Mev protection Every Sandwich protection feature and bump their slipage to like 50% just to be first so people like do it on like themselves will like go through the ux through the menus to go and do this so they will override the defaults um and this is like because uh the hot spot here is any single transaction that touches that thing changes the price because it's so

18. 9:14

    low liquidity so you can't predict what what that slippage is and you're willing to pay that extra and I think this is a very kind of difficult gnarly Mark you know micro Market structure problem that I don't know how to fix um I don't think there anybody really does but um it's going to take a whole bunch of research and kind of iteration there's some ideas um max who was like you know uh G gave up on ethereum and

19. 9:45

    join onza has has some pretty good ideas about how to potentially combination of priority ordering and replay plus the application being aware of that ordering and being able to move its price um relative to to like the execution to to kind of avoid some of the the worst case scenarios and if you can avoid worst case scenarios then in theory people providing liquidity can quote tighter spreads and and Things become better and move faster and stuff like this um so

20. 10:18

    the the bottom legs here are not even like it's not even a capacity or physical constraint on the network itself it is the like information light cone if you want to think think of it it's just locally the price changes by the first photown that touches it just and now everyone that thought the price was X is now it's completely different um if you if you if you can observe the price it already changed exactly that

21. 10:45

    this is like a true shoninger problem um I I I don't think there's any solutions even in traditional finance and most folks kind of point to like discret batog options and things like that as kind of better Market structures that exist um than than currently exist well I think the the problem is people want this Market structure right people like it's effectively like a gambling Market structure where you don't even know what you're going to get but yeah at least when um when we were

22. 11:19

    working on nfts um the the the the teams working on this have have outpaced my ability like my knowledge and understanding of of Finance um so but at the time many years ago when it was just nfds on salana there was similar problems where you'd have like people make a mint and it would do the most simplest mint just everything opens up and first come first serve on chain um so people would just Spam The Chain like for 30 minutes prior submitting bids that would fail so

23. 11:54

    they're willing to pay the fee repeatedly just for the higher probability chance of Landing first um and getting that thing and we rebuilt we built Dutch auctions we built every kind of auction you can think of and nobody would want to use this because it turns out people actually like the idea and the event like both the founders and the users want the spectacle of the sellout the can they they even like were proud of the congestion that they cost

24. 12:25

    yeah they want to see the congestion yeah because that that shows how how valuable that is yeah yep so there's some yeah some some weird things there that make it really really hard to change user behaviors in the application layer for to be easier for the chain um um I'm not sure how to address that like the best we can do is basically isolate them and now like NFD mints don't really touch the rest of the chain but you can still create a mint that's going to be

25. 12:59

    you know arguably a bad experience for the users but have a this is what the users want right yeah yeah I mean we we've been thinking about it a lot and kind of one of the uh like fundamental challenges of the kind of the way right now especially defi works is that yeah you you're submitting transaction without knowing what you're actually going to arrive to to right and uh there is a different approach where

26. 13:33

    you effectively first figure out what you're going to do right and then you already settle that commitment and so that's kind of this intent architecture that uh we've been working on and that that's closer to a like a central limit order book where you have effectively orders that can settle later that people have posted whenever the price crosses and stuff exactly yeah like in a way you

27. 13:58

    you I mean you decompose the central order book and you say hey we have like the blockchain serves as effective The Ledger for set for settlement and and asset control and then yeah you have some way of matching right this can be peer-to-peer like we can literally agree on something right now and this will be an OTC transaction uh and we settle that on on chain it can be like hey you know

28. 14:22

    who gives me the best price or it can be indeed an order book where you know uh orders are matched and the and and the the interesting thing is like you can have those kind of local and Global markets in in this structure supported at the same time because you can be matching with somebody who's you know here in San Francisco you know wherever 50 millisecond like latency and you know if you're willing to wait you can get quotes from uh from around the

29. 14:54

    world as well what is the tradeoff there with with intense uh I mean you do need you do need a more active uh kind of liquidity provider right on the other side so a like the basic like amm structure right where you just you know set it and forget doesn't work exactly like that now the cool thing is we now have also ability to run effectively autonomous like uh actors outside of blockchain right through TV I

30. 15:29

    mean there's some additional security assumptions that come in with that so is there a cost because you're a liquidity provider you have to go put up Capital you have to maintain yeah some some of the yeah so so it is very similar to how you would do it on uh centralized exchange or yeah or or order book yeah yeah so the the I mean this stuff will work for tokens with a ton of liquidity because the the problem is like you as a market

31. 16:03

    maker um you have to put you know you have a CH choice between your dollars being in t bills for the year for three months or buying this Capital that then you post at a position and when you buy the token that is like L liquidity yeah your your risk and volatility goes up so much that the Roi on that that to be has to be really really large and this is why I think that it it's

32. 16:37

    going to be kind of hard for that early stage Capital formation process to move to intense uh and Central limit order like clobs even so even like onchain with all the other available liquidity like a super optimized Central liit order book with like their own market makers basically like Phoenix can't actually serve the 40,000 whatever coins that are created every day because it you just can't afford to go buy the capital to to post it like you're going

33. 17:10

    to see that stuff working for some things but not for everything else right so just like everything else once the problem is complicated enough uh some of the some of the things will work somewh yeah I mean obviously decomposing it into yeah different stages effectively of tokens make sense I mean one thought we've we've been working on is effectively you can have a amm style Market maker that like when you launch a token

34. 17:40

    right you just put like effectively similar how you would put it in swap radium um type and and it effectively just is able to trade against all of those requests yeah people have built those for youum those actually think almost like radium's first design as they they sat right on top of serum what I remember it's pretty pretty interesting the near project that built the intents out um what's the goal of

35. 18:14

    that I mean so the intents from my perspective is is actually how not just crypto but also like Global Commerce layer right because you know you can also find people to do marketing right or you know move some steel or uh like you the intent itself can be a lot more generic than just you know crypto element so it really just depends on how you like how you verify the delivery and how do you like dispute if something

36. 18:46

    goes wrong so so kind of the way we look at this is intents are this like kind of interconnect that different actors right be that the AG Financial participants users consumers Etc can now uh do trade is this the primary way that you think you can unify cross Shar liquidity so so I mean we we already we on near itself we don't have the crush sh liquidity problems but the we actually by doing this we unified the cross chain liquidity right we have Bitcoin we have salana we have xrp we have Dodge you know um like more chain

37. 19:28

    of La chains as well and so that itself is also like serves this kind of unified market for and so I mean that that's why I'm like I'm curious because I mean that that's kind of been the pitch right for uh for Sal so I'm curious how how you see those two different approaches I mean like you see like a combination of like Wormhole will provide rappers for tokens and once uh once they exist

38. 19:59

    then you can effectively like in Jupiter you can the token doesn't even need to be fully wrapped yet when you buy it it can be wrapped on demand and stuff like that so you can you can build that build that similar functionality it's going to be largely depend on product and stuff like that the nice advantage of having everything in in one unified State machine is the um you have a lot of people experimenting with different markets and you have aggregators that

39. 20:29

    can pull all the liquidity from those markets together so functionally you kind of have very simple products you know you have a or radium or whatever they're they're building out those pieces and those markets are competing with any kind of intent based order book um and then you have an aggregator that can pull liquidity from all of them and they're both competing with the markets themselves but also providing them a

40. 20:56

    valuable order FL order flow and Stu stuff like that so you kind of have this really nice Loop that's happening on salana right now um that hopefully it's hard to disrupt and just keeps going and then I don't have to do anything and it's easy to know what to do for the L1 right you just make it cheaper faster and and and things just keeps getting better the I think the CH the advantage that that has is you have very yeah have very many different

41. 21:29

    points of user acquisition that are somewhat uh not entirely like competitive not not totally adversarial to each other um so even some of the marketplaces work together like pump and radium you would think why would why would pump want to use radium it's easier it's available right it's in the in the same ecosystem they didn't they just use their apis and it was much faster for them to launch

42. 22:02

    um I don't we haven't seen anything like that really spring up in Cross chain or cross L2 even um yeah I I think the because the communication itself is so complicated right and like you're kind of relying on Stitch work of of different solutions right now it's it's hard for someone to rely on like okay this as a team will just handle stuffff for me I mean it happens right just like at at way smaller scale if you

43. 22:34

    were so like if you were to go build an L1 with andrin solutions for all this stuff the problem is you just built another L1 they to go find apps to go get users uh and are those Solutions more than marginally better than what exist is also unknown um even if even if there's a bunch of inefficiencies in the curan kind of setups users ultimately just they have their wallet they hit the swap button and can them or whatever and

44. 23:06

    that that's good enough for large majority of them yeah or even simpler they have a telegram pot yeah it's just a telegram bot or soon it's just going to be a i bot that trades for you and uh you just uh just send it tweets and it's like figures out what you want yeah so maybe that's a good segue into uh some of the AI stuff I mean we we've seen a crazy amount of uh activity over over past uh weekend and kind of the market freaking out as well um curious kind of where where you've

45. 23:44

    Fallen this uh on a few of the different dimension that people are discussing around like National Security around the usage but also like in general interesting to hear your thesis on this I I think people are overhyped in the short term underhyped in the long term that's a cop out uh look ultimately AI is just a bunch of data it it it's just like some files on a hard drive um it's not in itself

46. 24:23

    autonomous um you have to have people to to maintain those systems and drive them forward even if you make it all smart contract based and and whatever there's still it's still ultimately just a a bunch of data that's automating a task but I think the tasks that it can automate are going to be kind of subtly very transformative um because you kind of you're lowering the differences between like the very easy example material

47. 24:59

    writer and I hate I hate writing I hate writing a long form it takes me a lot of D we all know yeah structure my thoughts but I can literally dump out what I how I think something works and then tell the AI write it for clarity and they'll understand me so AI is smart enough to understand my kind of like stream of Consciousness and write it out in a cohesive well planned well well written document and that limit

48. 25:31

    that totally reduces the my gap between me and Paul Graham or something somebody like that that is an excellent writer that is very thoughtful and spends a lot of time now obviously Paul Graham is still way ahead of me unlike structuring thought and being able to you know make good arguments and stuff but that Gap is dramatically closed it's not it's not like you know 100x and anymore maybe it's just twox uh I think you're underestimating

49. 26:04

    how your concise bangers on Twitter how valuable that is that's actually harder to do than long form writing um no I'm I'm uh it's easier because I'm I'm borrowing a lot of Concepts that everyone else is built up and I'm just gluing a few things together without having to go explain how all these things are interconnected and people already know like the Banger people already know what I'm saying they already had that thought I just get to

50. 26:36

    repeat it and they acknowledge it I think that's a very subtle but uh very big change because the way that like big changes move forward is actually through storytelling and ideas and and things like that and if you can concisely explain how big things need to change they'll actually happen so I'm pretty excited for that I think AI agents reviewing what the hell government does summarizing it explaining it people

51. 27:09

    dumping their their like reaction to that right and then having AI reconstruct it into sound actual good arguments and get feedback in that process is going to massively speed things up um in a in the US for sure I don't know about kind of more bureaucratic societies but I'm I'm pretty stok for what's going to happen in the United States yeah I mean I think people already like I I've actually met number of people even in like at the government or uh around government

52. 27:43

    whove actually being successful because they managed to like dump the new bill into CH GPT and effectively being able to ask questions and you know when you need like the the whole you know spending bill right is like hey we need to pass it tomorrow here's you know 500 pages go right like it's impossible to really analyze it but you know if you could just like hey you know tell me what are the like five most expensive items you know who added them explain why ET right you can start just

53. 28:15

    like tearing through that a lot quicker um yeah I think I mean I think what you're describing is like there's effectively few stages of from my perspective of kind of a adoption right and so that is like a co-pilot stage right this is where like you are driving but uh you know AI is like refining your thoughts or or helps you with like information gathering kind of basic analysis I think what I your point about like people

54. 28:48

    underestimating where this is going is when it does start making decisions because a lot of Our Lives is just a a set of small decisions and set of small markets or large markets right and so I'm I'm actually really excited when when we can start a floating s the decisions to AI because you can specify your preferences you can specify your kind of show how you've made decisions before and now it can effectively apply that

55. 29:20

    pattern going forward but you can scale that a lot wider uh and you know this can be on the government like for government right where like literally can vote on every Bill like every individual can vote on every bill just their eyes do in it like real time uh but also you know like which food to choose you know which drugs like you don't need FDA be essential authority to review everything and decide if it's healthy for everything right each individual can like review every single drug and decide if it's healthy for them

56. 29:52

    or not so things like that right but you know it scales kind of across the industries yeah the danger there is it kind of becomes like your your Google Maps where you stop memorizing City and you just like trust the map to tell you how how to get from point A to point B um you you were Mobile in the 90s or mobile enough to have your like built your own map inside your brain of where you

57. 30:24

    lived yeah like pre preap Quest so that was just normal part of life right you kind of like by driving everywhere you never I've never looked at a at a mapping thing while I for like the first five years of my driving uh life um so I wonder once people start delegating more and more tasks to this thing that is mostly correct most of the time do we give up on the muscle yeah yeah do do we

58. 30:55

    give up on the muscle that's probably the one danger the other danger that I joke about is that if an AI can infinitely entertain people if they can just constantly tell good jokes that's that's the the sneakiest way to attack Humanity that's called tiktock no no T inite feed of you know of entertainment Tic Tac is not as good as what uh ASI could do right in theory that's true yeah yeah I mean I

59. 31:34

    think I think there is an interesting question there which is I mean the like and and partially why I'm I'm very concerned that if if there's like a single for-profit company that just kind of you know presses on that like gets it right and just presses on that right then indeed we can just end up as like right just maximize the attention maximize everybody's like time spent on

60. 32:00

    this and uh it's a single company if it's a monopolis they don't actually need to do that because they're already capturing all the value the problem is is if there isn't a monopoly and there's competition and they will constantly tune it to be more and more entertaining well this is why we need it on the user we need something that's like on the user side and actually cares about your

61. 32:23

    like holistic you know wellbeing and is actually interested in you like not atrophying your decision- making muscle be you learning new things and kind of you know expanding like we kind of need to tune for that that's like asking people to exercise for help I mean there is people who are exercising right there there is there is intrinsic motivation to do that a small percentage yes yeah well let's let's get let's get

62. 32:53

    everyone exercising too yeah no I I I mean I think that there is interesting kind of tensions here which is indeed if it's like a pure capitalism that indeed like you have kind of motivation for um for everyone to effectively like try to steal your attention and kind of create more addictive products so I do actually think I mean where crypto comes in is the whole point of crypto is not just pure capitalism it's actually like having a CO ownership in the system

63. 33:26

    which you know should grow in value but it it also doesn't need to like at some point right it can achieve some state where you know everybody's kind of participating and um and so there is an interesting argument that a crypto system can actually be more user driven than a for-profit company would be why wouldn't those those users attentions be eaten up by a more aggressive capital is driven competitor for the same functionality it's like send mail right

64. 34:06

    you run your own sendmail server and then Gmail made it made it extremely easy to not do that because they earn a profit from you not doing that yeah by showing you ads and showing you other stuff well I mean it needs to be it needs to be as easy it needs to be competitive on the product side so then it's also turning up the entertainment button the AI the yeah release Ender fins buttton yeah y yeah I think we're

65. 34:48

    screwed we are the reason why we're live during this age is because this is the transformation from that asymptotic into Singularity St and this would be the most interesting thing to simulate post Singularity to see if there's any other way that Humanity could have like could have went yeah could have could have done made any other choices we're just one of the millions of iterations making the same mistakes I mean I don't know if those are mistakes like the thing is I think

66. 35:25

    that this evolution is pretty natur contr right as kind of comp Computing is done like everything is kind of pointing in this direction so so I think it's indeed is a question like what is that final state where the Computing and human are effectively on par or Computing is beyond that and like it at some point I mean like I I agree with you that like there's intermediate step which is this kind of capitalism you know let's just take all

67. 35:58

    the users attention and my point is we kind of in that right now that that's actually like we kind of have like you know few of those products which are effectively right now competing for users attention and trying to maximize it they're not AI driven though they're still I mean they they very much like Tik Tok is just an algorithm it's like a it's like a sure but humans that you're

68. 36:21

    still have people competing to provide the content yeah I mean for now so that that's the problem like next two years once that goes away and you it's all purely AI generated then I mean our X feeds are probably you know half AI generated at this point that's why they keep getting better and better yeah but uh yeah but I think I think like the again to me the next evolution where we are um like where we see like a a real transformation not in

69. 37:03

    entertainment but like in kind of how we operate dayto day I do think that that will also change how we perceive entertainment and and also like in some ways canot can undermine in the capitalism systems that are generating those kind of entertainment apps because like if you just buy a phone and phone effectively has your assistant and it does everything including generate entertainment for you right but you know if if if you just like get stuck in that and you're not

70. 37:33

    going to like like that also not not economical so I think there there's interesting systems that can be built out of that maybe maybe actually an interesting topic given uh the salana mobile like how you see the evolution of of devices given AI as well I think cost to run AI is going to continue to drop so it'll be somewhat commoditized and the big winners are going to be kind of meta and

71. 38:07

    apple um because meta has this massive data and and user advantage and Apple has the device Advantage so they kind of have these natural modes where I think it's going to be actually kind of hard for open AI or whoever else to really create like a sticky I the thing is like I actually think what deep seek shown is that the data mode doesn't really exist like I think that at least on the AI side itself like a lot of I don't think for training it's

72. 38:42

    it's more for the just the users interacting like you have the network effects of users talking to each other um that's hard to move and you can't like move all the Facebook's users and their interactions and post into chat GPT so chat GPT is kind of single player Facebook is not single player by default and that has a huge advantage in terms of deploying AI services and and like things like that for the consumer um so

73. 39:16

    if you have like that if you have a large consumer base that is has Network effects between them that those are it's a good spot to be um similarly for like because of just how well their devices are plugged into every aspect of your life it's just hard to remove them so it's easy for them to take the cheapest implementation of AI they don't care if it comes from chat GPT or from Deep stack like the model is cooked it's like

74. 39:48

    open source they'll just plug it into everywhere it makes sense yeah I mean Apple for sure is is is definitely an interesting position to capitalize on that uh we do we do need an open version though that is uh yeah so you guys building the trillion parameter model yeah that's that's one of the things I mean that that's that's one is more about like right now effectively everything that's even open like what we call open source is actually not open source right it's open parameters models

75. 40:21

    but the source code is not known like we don't know what data went in we don't know what like post training was used uh and so given all of these models are just data right if you don't know the data you don't know the like what actually it's you know uh able to do how reasons and and it's important because there's actually ways to do sleeper agents so like I think everybody you know I mean and like deep seek is obviously very exciting but like there

76. 40:52

    is known ways to effectively train a model to do things based on triggers including like based on years so it can like only you know after some date on the in specific context Etc which which really hard to hard to catch unless you've seen and have verified way of of evaluating the what went in how it was trained and so that that's kind of the idea like how do we train an actual open source model like at sufficient scale obviously us can distill it to smaller

77. 41:23

    models so uh for actual usage but that everybody can inspect the all the inputs and kind of understand like even prove that like hey it contains this data it doesn't contain his data I think open training will actually be somewhat reasonable because the blood of Hardware deployment is going to start to dep you know start depreciating and you can just use previous gen stuff to catch up um those

78. 41:54

    data centers are going to need customers in two years right like so um I think I I think it's somewhat doable I mean even now there's actually like there's a ton of like two to 4,000 GPU clusters available like h100 clusters because people bought a lot of them but uh it's not it's not as easy to utilize like mediumsized cluster as people think yeah I mean there's there's definitely a bunch of interesting research on how to do kind of distributed training as well uh which show starts to show interesting like

79. 42:31

    promising results around um just like how much data you need to uh send around yeah I mean I think the the open version though like we do need a device that's like really everybody can rely on that it's not like trying to trying to do all of those things with you right and maximiz you in their fin release uh something that like people can actually truly own um like I am I am excited for you know to see that uh to continue

80. 43:08

    happening yep yeah me too what's uh I guess like what's for this year what what are you the most excited about I mean in salonic system and web in general the usual stuff stable coin ad option I think there's going to be a trillion dollars worth of stable coins and and 5 years so figuring out how to get traction there is probably maybe the most important thing um and it's boring it's payments and payments is hard but I

81. 43:40

    think it's that's probably the most transformative thing that that crypto will accomplish um the other one um obviously Seeker launching so that that's getting close uh excited for that and then trying to build a ecosystem around it is a SLO but also exciting um and fire dancer is pretty close uh I don't want to jinx them but they're making a ton of progress I think full the full Noe is

82. 44:18

    staying up between uh like an invalid like a block cash mismatch Bank cash mismatch like for week now um so the their like errors are getting more and more rare basically feels like in the final stretch do you feel like everybody will switch to fire Daner like the adoption on the valer sites um I mean it will definitely get more than 33% adoption once it's out then you need somewhere between 33 and

83. 44:50

    66% for the performance um it's just for safety right because you want them to Halt the network if they diverge if there's a bug yeah yeah well I mean also like for median performance as well to improve you kind of need like is pretty fast so actually the challenges with improving performance of the network is more to do with um like bookkeeping and accounting and the svm of of all the resource usage

84. 45:20

    um the make like it it's basically like um the it's really really hard to accurately measure uh how compute is used and and the cost of that usage in a multi-threaded VM there's just a lot of interplay between resources so you're kind of doing your best effort to be wor you know prevent the worst case from happening if you double the block space you're allow you're effectively giving the attacker more leverage so tightening

85. 45:56

    those measurements and making it more robust and stuff that's maybe the the biggest challenge U but it's it's just for yeah I mean we we've I mean same same thing been dealing with a lot of that from early on just like running simulations at different scenarios right you kind of need to understand like cach misses the across you know five layers of different systems like how are they

86. 46:22

    affecting everything what is the worst case scenario what's the best case scenario I I definitely think I mean stable coin adoption getting getting more applications getting more just people on chain as well right um even if they don't know about this right they're just like hidden behind you know some even fintech application or uh some of this applications who just using blockchain on the back end I mean I think the the SpaceX example was

87. 46:51

    really cool right like they needed to receive money from Brazil they're like oh yeah stable coin is actually the the easiest way to move it around fastest too yep so yeah so I think that that's very exciting and and I think now now yeah there was like political shift it's not it's not a taboo anymore so it it'll be way easier and uh yeah I think in four years we're we're definitely going

88. 47:16

    to have so so much of this already like ingrained into into Financial system uh but even you're right you're right that like even past two years when they were kind of to kill Krypto it's like you you can't really do that like it's the the analogy I've heard from from China was like you know crypto is like water it doesn't it doesn't matter how many stones you put in front of it

89. 47:39

    right it's just finds its way so uh I think I think every every government have learned that already so now it's now it's really like about adopting it and actually like you know using this water themsel to carry them forward I appreciate you joining in uh great conversation there's a lot uh there's a lot to do there's a lot to learn but also worth appreciating that are you gonna move back to the States I'm actually in San Francisco right now oh

90. 48:09

    right on yeah uh yeah I'm here for a few months and then we'll see um but yeah I mean definitely definitely the cent of mass is here now so it's it's it's important to uh kind of keep building here awesome yeah I'm I'm actually like two blocks away from from the you know from where our offices were oh cool yeah our offices were actually like across the street from each other when we were

91. 48:38

    starting in 1819 so uh we would me to for coffee yeah yeah so but yeah I just wanted to say yeah like it's been it's been a long road since then but yeah I I mean obviously gu guys done amazingly well um and uh it's uh yeah it's exciting to see kind of progress across the board thank you yeah likewise and good to see you yeah good to see you