L2s Wont Scale Ethereum | Multicoin

Transcript

this episode of Empire is brought to you by quick node quick node is an end-to-end blockchain development platform that makes building web 3 apps super easy no matter what you want to build you can effortly develop any application by leveraging their elastic apis go to quicknote.com use codeempire you'll get a free month on their feature back build plan that's right go to quicknode.com you'll get a free month to start playing around you'll hear more about quick node later in the show this episode is brought to you by synthetics the liquidity layer for D5 derivatives with synthetics V3 any protocol can now tap into synthetics liquidity to bootstrap derivatives markets you'll hear more about synthetics later in the show

welcome everybody to another episode of Empire we have two very special guests we have Kyle and Tushar from multi-coin Capital um I've got no uh both of them as an investor in multi-coin I vividly remember reading one of Kyle's first blog posts and thought okay this is interesting I don't necessarily agree with everything here but it struck me as um pretty interesting takes and that led me to reach out to them and you know I followed their Journey um over the years um and uh really uh I think the combination about Tushar and Kyle is very interesting um and so anyways we this should be a great episode um recapping everything that's been going on lately and also uh some of

their thesises as they see the space going forward so Kyle Tushar welcome back to Empire excited to have you guys on thanks for having us good to be back on the show guys I want to spend a lot of this podcast think like thinking about kind of updated thesis around uh Solana like other l1s different ecosystems things that you guys are excited about right now but I think there's this kind of like elephant in the room about you know multi-coin was kind of riding higher than any other fund out there I think you guys avoided some of the like mid 2022 snafus like Terra Luna and stuff like that but then you did get caught up with some of the FTX stuff and because FTX was so involved with Solana you guys got hit kind of twice with Solana stuff and FTX stuff and haven't really heard

from you guys since then so did just want to like address the elephant in the room and I think a good starting point is like how did you get so intertwined with FTX and like what was the relationship there what like what did that look like shredders I'll start and then sure feel free to to add on top of this so you know we we Sam ditched us to the FTP seed round and I think more of April 2019 we did not invest uh at that time uh but we certainly watched FTX kind of grow it's still I think the exchange launched in May or June of 2019 and we kind of watched it grow in those early days and by the end of 2019 we started to think hey this this exchange is pretty interesting um Solana then launched blockchain three

months later in March April of 2020 uh and right around that same time uh D5 kind of started to get interesting you started to see governance go live a compound liquidity Mighty would launch in March excuse me in June of 2020 now it'll like kick off D5 summer uh at the time I remember Sam was tweeting pretty aggressively about defy and I was like pretty intrigued with with what they were building at FTX and it was pretty clear that they wanted to do something with D5 um so reached out to Sam at that time it was like hey uh what are you doing like you're obviously trying to do something and he's like yeah I'm trying to build something yeah and we just can't make it work and I was like you should check out this a lot of thing uh at the time slot

it was three months old no one had heard of it and uh no one took it particularly seriously that was kind of the start of our formal relationship with with Sam uh we did invest in serum um which which launched like 30 days later um and then as FTX kind of grew in prominence over the course of 2021 they started raising money uh we did invest in ftx.com and in STX us we did not lead any of the investment rounds in FTX uh those were led by some larger investors but we were participants in both of those rounds so when do you guys actually know that things were going south because things escalated fairly quickly you know when when CZ sort of tweeted um and that they're gonna sell them onto

the FDT like it was sort of a timeline of 72 hours things really escalated when did you guys actually know that things were going down in FDX and how did you kind of operate in like walk us through kind of that timeline yeah it was during break point so we were in Lisbon uh and kind of watching the events play out uh and also watching the community sentiment as things played out right it was like a major topic of conversation as you might imagine um honestly we didn't think there was going to be like this level of insolvency or this big of a problem uh especially given the very large amount

of equity Capital that FTX had raised that should serve as a buffer and in addition to that this was a profitable business that made over a billion dollars in Revenue in 2021 so there was quite a bit of money there so we thought okay there might be like a duration mismatch um or you know their capital is tied up in margin positions and it's going to take them some time to unwind it and that's why you know this is happening we didn't think it was going to be an insolvency problem as it appears to be hmm was there any so if I'm trying to remember the dates it was like break point was November four five six I think

it was the coindesk report by Ian Allison was November 2nd was there any like hesitation or like skepticism when I don't know what's what's another Moment Like Brett Harrison for example resigning in late September was there any sort of like hmm like Brett good guy seems like something fishy is going on or like that just that doesn't really cross your mind when things are going so well um certain things like that across the mind and we were investors in in ftxus and so we we had a good relationship with Brett we still do to this day uh and we called actually Brett called us um you know after he had little sand here before it was made public to let the other STX who has investors no that was kind of the right thing to do

um and he explained he's like look like I've been here for a little while uh I've caught the startup bug I want to go launch startup uh yeah he's obviously I don't think we I don't know if he's announced the name for it but he's building some sort of kind of D5 Prime brokerage type tools something in that in that Arena and uh we were like look appreciate you being forthcoming and candidate about that with us uh we'd like to take a look at your new startup whenever whenever you launch it uh we did not end up investing in it but we always had a good productive relationship with him and he never alleged or alluded to anything being wrong in me to respected anything uh it was just kind of like a hey he's been he left trading for our King to kind of

start up land that was like got to start a thug which pretty pretty reasonable thing to happen so I think it's fair to say that to show what you said which is I think a lot of people maybe have expected that there was a duration mismatch but not the level of just outright fraud that was going on and just the loss that was being carried by Almeda like as you look back were there like what was your impression of Sam like throughout all this time and and like what are the things that you guys got wrong in this whole process uh so my impression of Sam is he's one of the most hyper intelligent hyper motivated people that I've ever met

um and you know honestly I didn't buy this whole media tour after the fact on like oh I didn't know what I was getting myself into right like this guy talked to several of the most sophisticated investors in the world and got them to invest including like Sovereign wealth funds and you know just like giant asset managers like the you know some of the most storied Venture Capital funds in the world and they all loved him because he told a really good story he was hyper aggressive and he was focused on addressing an enormous Market so I think yeah I I think everyone's perception of him was pretty similar right like this

guy was kind of like a superhuman um right but what I think was underlying that that that we missed was the complete lack of governance controls that anyone placed on him the complete lack of transparency into the internal operations now look when the CEO of the business goes to the CTO and says please write this back door into the code such that it doesn't alert the Auditors when I move this money uh there's very very little that can be done from an oversight perspective to prevent that from happening but you can have some processes in place uh

right if you're checking to see where the money is if there's an up-to-date balance sheet always presented if you operate the exchange more like a D5 protocol where everything is auditable in real time uh and I'm hopeful that we as an industry will learn from this and move exchanges to operate more like D5 protocols with real-time transparency rather than this periodic check-in process we have the future of Finance is not quarterly reports that's you know that's an anachronistic 20th century thing the future of finances real-time visibility into everything that's happening right now yeah and I think that's probably the one

thing that could have prevented this yeah so just round up the FTX stuff I know we there's other stuff we want to cover but the last thing I want to touch on is what happens to the funds that are stuck there because I mean you guys had meaningful Capital there as did other folks you start to withdraw you take some out there's still some there left um so what is kind of what happens to those funds I mean the bankruptcy trustee will try to liquidate all the assets that they have um some of which it appears they have already done based on some of the announcements that that they have made some of which they have not done um and then everyone gets paid back uh

they're a big open question right now is whether there will be clawbacks for the funds that were withdrawn right before the bankruptcy but the question is like how do you enforce those especially with International customers and uh like it's just really messy uh right and you don't want to treat people differently you don't want to say well actually only the Americans have to send the money back and everyone else like you know that creates lots of weird problems uh so I think that's the biggest uncertainty and then the next thing is just like what happens to the Investments that they made for example they LED you know this investment in proper Ai and we've seen this explosion in AI market caps and enthusiasm and uh

you know who knows how much that's going to be worth or or when that's old so you know we are optimistic that there will be some recovery value for the creditors uh however it seems overly optimistic to assume that it's going to be you know 100 that that seems unlikely well particularly considering that lawyers are you know I think that I saw a note today that they're up to date like they've taken 30 million of these so far and you know I think if we look at the Madoff case like these guys will probably take meaningful cents on the dollar that is recovered so um you know that's a shame Kyle was this the uh I mean you you and two sharp have been in crypto for a

while now but I mean obviously we're all used to like big ups and downs um both like in the portfolio but also mentally was this uh like the hardest mental thing to go through uh yeah it definitely was the hardest mental thing you know 2018 that that cycle uh like crypto wasn't taken seriously by the outside world like there was I wouldn't say there was like a real Spotlight on crypto and certainly not a spotlight on us in any capacity um and having the year we added 2021 then like everyone being like Oh yeah look at these these suckers these idiots you know whatever getting getting dragged out on all this it was a pretty shitty feeling and then obviously people are angry people have questions people

alleging we were involved it's just like just uh you know uh the conflicts of just a lot of mostly negativity and then it's not legitimate questions uh it's just a very frustrating time and like you know we told the whole team like hey y'all are gonna be fine everyone's gonna keep their jobs Keep Your Head Strong stick through it like yeah well we'll get through this and and we did we haven't let anyone off and uh Martinez continues to go strong and now you know with most of the crisis behind us we can now Focus again on what we're supposed to do which is find amazing Investments nice um yeah I mean I talked to some Port codes and some folks at mult like people at multi-coin and saw you guys just had

a couple big promotions and like I think the team team seemed stoked right now portfolio company said you guys treated them really well it seems like maybe the the toughest place would have been with the LPS I saw some like Angry Twitter threads and uh you know there was like a couple stories about the fun being down a lot and like clawbacks and like how do can you just tell us what happened with like the LPS and obviously I'm sure 95 of them were like fine and understanding but I'm sure they're outlier five percent cases where there's some some angry folks what like what was it like dealing with them look uh whenever you have events like this like there are going to be angry people and people react to anger in very different

ways we are happy to have a strong base a very Diversified and long-term oriented LPS many of whom have been with us through multiple bear cycles and they know the volatility can be pretty extreme you know we are not the risk off part of anyone's portfolio yeah that's not the the purpose that we serve in any way and you know everyone knew the risks uh which were pretty clearly disclosed but look we had a bad year like we acknowledge that we own that and we learned from those mistakes um that being said the only way to earn excess returns as an investor the only

way is to be contrarian and be right if you are contrarian and wrong you don't earn excess returns if you are with the herd and right you just earn market returns uh and obviously if you're with the heard and wrong that's that's pretty bad right so the only way to earn excess returns is to be contrarian and right um and you know that's what we're focused on uh we're focusing on capturing value across Market Cycles not within Market cycles and while the hedge fund is uh investing in liquid assets it's not investing with a six-month time horizon or a one year time Horizon it's investing with a longer term time Horizon and we like to you know joke internally uh a lot of disagreements come down to a function of time Horizon

yeah on that point uh I do want to touch on sort of position sizing at what point Solana becomes a disproportionate size of your fund and it took a while um right what is what's it like like managing that position in hindsight should you have sold a chunk of that um let's start there and then also want to just generally touch it on Solana how you perceive that ecosystem now yeah I can go first on this and hand it over to Kyle I would say look the people who are going to complain uh are gonna complain either way if we had sold everything in 2021 they would have complained that we sold everything in 2021. if we didn't they were going to

say oh look you're you're dumb for not selling everything in 2021 right uh so we try not to pay too much attention to that I would say that you know we continue to have extremely high conviction that swanna represents the right design choices for a scalable blockchain and we like that much better than the L2 designs that we're seeing out there and we're focused on the long term right like if you invested in eth early on and then you saw it go from 1400 bucks to 80 bucks in the last cycle uh that felt pretty terrible uh but then if you had the conviction to hold on you got to see

ethad you know what's it at right now like 2000 um and you know much higher than that in 2021. so you just have to be willing to uh accept that prices are set on the margin and that humans are really emotional and if you want to outperform you have to do things that other people do not do yeah over indexing on cost basis is uh it's like a bias in your logical reasoning it's almost like sunk cost fallacy right everyone's familiar with some cost fallacy the money that you already spent like you've already spent it that should not be a part of your decision moving forward you should look

at the facts and circumstances today the same thing is true of looking at cost basis right like the fact that it was priced at X you know however much time ago is not necessarily relevant to what is it worth today you have to look at the price today you have to look at what you think it's worth today and they make a decision based on that and you can't beat yourself up too much uh right because if you let me trade in hindsight I would be the world's first trillionaire right uh and it's very easy to trade in hindsight but you just have to you know always be looking forward from the moment that you're in uh and I think that's the best way to be as rational as possible in a market like

this so talk about fundamentals like and circumstances and kind of updating your thesis how much of the FDX impact and we hadn't totally right after kind of the FTX stuff happened and we asked him a question we'll ask you as well is does that does FTX and disassociation and the perception of the market that impact Solana is kind of BD efforts which at the time were going very well there are some teams may have left other teams you know it's a competitive kind of competing L1 ecosystem they can go to Aptos it can go to Avalanche and go to elsewhere right an L2 how much of FTX do you think impacts Solana um viability or just the willingness of projects to kind of work there and you

know obviously the defy ecosystem in Solano took a big hit um and I am curious to just take your pulse on on Solana the ecosystem today and how you see it going forward particularly in this competitive kind of L1 um environment um I guess I'll start here there's this perception that like FTX was acting as the BD arm of Lana or something or kind of some other hand wavy-ish statements um and I don't think those are really rooted a reality um the most important thing FTX did first a lot of was built Sarah and which was definitely legitimate at the time they did it in terms of like the effort obviously they shipped the code and it

was open and then other developers used it uh that happened in 2020 mid summer 2020 over uh you know next gal a year and a half two years FTX did a ton of stuff all over they were investing in exchanges like IEX and other core C5 infrastructure things like uh Robin Hood and stock twits like they were making all these plays they are pursuing stuff with Aptos and uh and sweet the missed in Labs uh they were investing in much of an nft projects and NFC exchange that was essentialized custodial nft exchange um FTX was very much acting in their own best interest and you can follow their their dollars and like see where they're you know put in their money um so this kind of uh Narrative of like

FTX was like supporting Solana I just like I don't understand the claim both in human capital resources or financial resources um they were doing stuff that was in their own self-interest but like that's just they were doing that you know that's capitalism right so um we just never really agree to that characterization um and obviously narratives are narrative hey perception is reality so like yeah there's some element of that and hey memes are a thing of crypto and people say oh my God it's a lot has problems now I've been that kind of self-reinforcingly uh earned as a self-efficient prophecy I guess it was problems now uh but initially those things go away I mean narratives are narratives that they favor pineless this intellectual substance kind of matters

over the long term uh uh I'll pause there to show you an additive thing um I add something which is I was generally on this topic of BD for blockchains right people who have been talking a lot about this and I have some maybe controversial opinions on it which is I think BD for blockchains is overrated uh that's too much of a top-down approach I don't think you're going to achieve like the next big breakthrough in web 3 by getting a web 2 company that's web 2 to their DNA to come and build or some other

centralized company to come and build on web3 and you're actually seeing that now in the depths of the bear market like that's the thing that gets dropped right like Disney laid off their entire web3 team because you know they're cost cutting and they're like uh this wasn't core to us anyway uh so I think the right kind of quote-unquote Business Development for blockchains is really the bottoms up kind it's the hackathons it's getting people who are crypto native who are building net new things that have web free in their DNA and they're not going to scatter uh in their Market because one thing I can promise you is this won't be the last bear Market there will be another one uh you know after the

next bull market it's going to happen again uh and I think you want teams with staying power through that uh rather than flighty big corporations all right quick break from the show there is this kind of overused cliche saying in crypto but it's true bear markets were building and everyone tells you that everyone knows it what people don't know is that if you're building apps in crypto and building apps in web3 without using quick node you are building on hard mode so quick note is is this amazing blockchain development platform it reduces costs streamlines the time to market for your app and it offers consistent performance at scale for folks that have built apps you will know that there are a couple key points

here one quick node offers unlimited endpoints across 18 different chains and 35 different networks they have response times that are two and a half times faster than any of their competitors 99.99 uptime and a dedicated 24 7 customer support team if you've been listening to Empire for a while you might know that I am no Giga brain developer but I do know a lot of devs and a lot of great product teams at other places so when I see coinbase and Twitter and Adobe and openc and doing analytics all leveraging and trusting quickness node to power their business that's when we get excited and that's when we want to partner with them the best solution for any leading crypto and web 3 company that is seeking an

end-to-end blockchain development platform right out of the box so my message to you get off hard mode let quick note handle the blockchain infrastructure let quick note handle the security let quick note handle the performance while you focus on building beautiful products for your users visit quicknode.com super easy you can use code Empire you'll get a free month on their build plan so don't forget to use codeempire Santi and I gotta get credit for this one so they know that we sent you and you will get a first month free hope you guys enjoy it all right folks it is time to talk about one of my and a lot of your favorite D5 protocols synthetics synthetics has been pushing the limit in defy Innovation since 2017 and has just started its most

exciting transition yet with synthetics V3 with synthetics V3 any protocol can now tap into synthetics liquidity to bootstrap derivatives markets the transition has already started with synthetics perps synthetics perps Taps into synthetics's liquidity layer and is a new primitive that developers can leverage to launch defy derivatives the perps product has been going incredibly well so far hopefully you've seen it it's had some great traction hitting 500 million in Daily volume this March we know that liquidity rules defy and synthetics is becoming the modular liquidity layer for D5 derivatives as a Trader you can trade synthetics perps with low fees in over

20 different markets at Quinta dot IO decentrics.com and polynomial dot Phi and this opportunity set keeps growing with 10 new partners in the pipeline ready to launch Integrations on top of synthetics including front ends Structured Products and institutional offerings the team gave me a sneak peek of all this stuff it's really cool would really recommend you check out synthetics.io forward slash perps to learn more and if you're looking to build on synthetics pop into their Discord server reach out to the team directly make sure to tell them that Empire and Santi and Yano sent you again synthetics.io4ps you can also hop into the Discord server and reach out to the team directly how do you balance um

like signaling conviction behind your bets which you guys oftentimes Love sometimes like polarizing takes um while ensuring you don't miss the next wave um I'm gonna push back on this characterization like uh do we write clear Theses on our blog like yeah obviously um but like that's not what makes a good investor a good investor um all the entrepreneurs feel good about their VC right nice things about them but like that's not what makes us good at our jobs what makes us good at what we do is like we help them with their problems uh and that's like hey guys like you just put on whatever this

marketing whatever and like it's a here's a problem with it and like you need to fix this or hey there's a strategy you know over the next 12 24 months like here's the four ways whatever this Market plays out like which which one of these are you playing for and why what about the trade-offs like let's have you think through those discussions you know I think our core value prop isn't is in that stuff most of that doesn't make its way out of the Twitter or outdoor blog it's just kind of inherently like private conversation but like that that's what makes a good investor are going to Nest or not writing blog posts uh so I I think that's important to like just characterize like what what is our what do we do and what does it matter um in terms of like the fall part of

your question of like you know if we make invest in a does that risk us missing something in the next cycle and that that is always true across all Investments and across all investment firms and all Investments um because of like entrepreneurs being scared for competitive reasons or whatever which is a legitimate concern people can have um and like that that's just inherent risk to the business right I mean there's nothing no real way around it um we do our best to be open-minded and to look at new things and explore and like push the ecosystem forward I think probably the thing we've done that is like the most generalized solution to that problem is just make it very clear to people that we will do weird new

things um we were the first one in graph we were the first money in art we were the first money in uh what's it called improve of physical work and in helium and high map all that stuff and we've continued to Signal time and time again like we will go do things that on the surface feel pretty weird uh and like in our opinion it absolutely matters because like the next entrepreneur was saying the next weird thing we want them to see that we have a tribe record of being open-minded and trying weirdest stuff yeah I guess let me so okay so I completely agree with that that's I mean yeah that was well said my maybe where I was going with that though is like I think there's a lot of funds that were also long on Solana um or that are long on Solana but like

they could kind of hedge their like I was looking at who invested an arbitrary and optimism and like a lot of the firms that were long Solana also like they you know invested in either arbitrary or optimism I don't think you guys did either of those rounds and I'm like you almost this is probably not true I mean I'm curious if you think this is true but like you almost couldn't have invested in in l2s because it's so antithetical to like your Global State thesis and for another fund that like might have a similar thesis to you they can kind of hedge and invest in arbitrum or optimism because like they're not as vocal about their Solana thesis yeah you have to be contrarian and you have to be right that's the only way to make access returns and if you if you end up

investing in every single platform what you're really offering is an index product you specifically cannot earn excess returns like you know you do the math and it is not possible to do that so actually what I see there is an inherent form of lpgp conflict right the LPS want you to go earn excess returns they they're not investing in an index fund for an actively managed fund uh but GPS hedging their bets uh you know have a different set of incentives sometimes from the LPS so we really try to focus on uh you know doing what we think is the right thing for our LPS and trying

to earn excess returns above what index or the market is going to earn yeah how do you think about on the point Fun Size and structure you guys went from hedge fund you then subsequently launch a venture fund and then as AUM grows naturally Your Capacity to earn access return diminishes and then so you know a lot of these funds and that went from 50 million 100 million to billions it's a very different strategy I would argue uh now the market has expanded substantially but you could make the point that a lot of these funds are just tracking the index right um so how have you managed those kind of decisions internally

yeah it's a it's a trade-off actually look there are some ways in which having a bigger fund makes it harder to earn excess returns like you said you have more money to deploy and that means you have to either pay higher valuations or you have to do more deals that didn't end up working uh right and that makes it harder but at the same time having a larger fund allows you to like make opportunities or be involved in opportunities that you couldn't otherwise do right uh if you've got a 20 million dollar fund and like I know what that feels like because we started with that right like you cannot lead around uh for uh you know like a a larger round right you you just don't have the capital to do that you can't

price it you can't help set the terms you can't influence the project to uh have the type of token design or other elements that you think are good so you know I think it cuts both ways right like size can be it can make it harder to deploy that amount of capital but it can also create optionality for you that is quite valuable so you have to kind of walk the tight rope and stay in the middle there you don't want necessarily to be the biggest uh because that's uh you know probably too far on the Spectrum but you also don't want to be the smallest uh you know I think just like Goldilocks somewhere in the middle is just right I think the probably the most valuable thing that a venture investor can do for any entrepreneur is not the money it's

not the connections um it's telling them that they're wrong without something that ends up really mattering and you know if you're an entrepreneur who doesn't prioritize that I think you have a lower chance of success than a similar entrepreneur who does prioritize that so I kind of like that selection effect and you know what I would tell entrepreneurs is like go talk to our portfolio companies uh right like that's where you should build your assessment of if you want to talk to us not based on Twitter not based on you know political affiliations or something like that like that's that's just not a pragmatic way to look at the world and you know you're just

not being fully logical turn up so um you talk about sort of Market size and how you think about that and you're we're now at that stage where a lot of people from the outside in even internally are questioning what is the actual use case of this industry um not too dissimilar from the internet back in the day but you know nonetheless people are overly critical now maybe tell us recent Investments you made or just thesis that you've had in terms of the killer use cases that you're seeing um in you know I've heard you talk about airdrops and the extent of uh that could be useful to go to market but that only lasts so long if you don't deliver a good product

people are not using it then it doesn't really matter uh it only gets you so far so so walk us through kind of your thesis on what are the killer apps um that you see over the next couple years um I'll highlight one area I'm spending a lot of time on right now and I'm actually I'm writing blog posts about this uh at the current moment um and that's that's AI and crypto um if you've been following me on Twitter the last few months I've been talking a lot about AI uh not because multipoint pivoting because we're launching an AI sign or something that that's ridiculous uh most of the like intellectually very interesting um and then most importantly because I think there's going to be some big

winners at the intersection of AI and crypto and so I want to capitalize those opportunities um in the existing multi-coin funds uh I've identified just three areas at the intersection of AI and crypto that I think make a lot of sense um and so those three are the first is basically Airbnb for your GPU um there is an extreme shortage of gpus today uh all the major Cloud providers uh I don't think this shortage will be resolved anytime soon um the obvious solution is like hey there are a bunch of graphics cards people's computers all over the world they're just not accessible in Cloud environments today so I wanted to be

nice if you could leverage these these gpus uh to do training and to do inference or these large compute jobs um we've been thinking about this problem for a long time this idea has been floating around in crypto circles since like 2012. like the very first narration of this was proof of useful work when people were like still in the Bitcoin saying the vine as they were like oh man proof of work is great if only we could make the hashes useful uh I'm like that was like the first iteration of that was was GPU workloads so you know I got into Krypton 2016 and that was being talked about I'll have to go a lot and lots of that time an eye exact we've been following a bunch of those a bunch of them the new crop

launched in 2018 uh we looked at a bunch of those we didn't invest in any of them uh the first investment we made in the Airbnb for your GPU you know sector broadly as render token which we invested in in May of 2021 so roughly two years ago uh and uh we invested in render you know for the year but the timely visitor render we were not thinking about machine learning workflow is specifically although we knew that that was like an opportunity uh but the render Network I think the name would suggest is focused on uh 3D rendering jobs for like 3D designers and uh that that we we foresaw that hey that grows big enough though eventually we

would ml obviously stable diffusion hit in August of 2022 and then um December of 2022 and uh it's pretty clear that like you can repurpose those render Network machines for things that are not 3D rendering and so you know I think the Market's largely picked up on this and a lot of people are discussing the opportunity for uh the render Network in the context of of uh training and inference um and the render team has they've updated the network and support table diffusion jobs um and they they signal they're gonna do a lot more stuff in AI jewel is the creator of uh run the network I believe twigled this morning actually adopt about this exact thing um there's more teams working in that sector than just renderer we've made one

other investment that what I'll lose to call Airbnb for your GPU um that design space there is not going to be winner thick all there's going to be uh subsequence there with different focuses but that's like obvious area number one at the intersection of AI and crypto uh section number two um is uh RL HF with token incentives so this is reinforcement learning with human feedback um if you follow we have discussions around chat gbt open AI is pretty open about they do a lot of RL HF um in trading the models um the most simplistic way to think about role Jeff is whenever you use chat DBT at the end of response there's up

and a down arrow or sorry thumbs up and a thumbs down uh and you can tell the model how it's doing uh there's a lot more sophisticated ways to do rohf than just thumbs up and thumbs down um I'm not familiar with the interesting of how they do the the training internally but they have large teams it's pretty well documented they have large teams of people who aren't providing sophisticated feedback to the developers and to the model itself directly um and a function of all HF I think there's a pretty interesting opportunity to use token incentives uh to facilitate large-scale rlhf or instead of convertible high demand models I'm not particularly convinced that this is going to be useful for a general live

model like child ABT or Bard or Microsoft Sydney or any of these but as you get into vertical domains things like medicine things like law things like Finance um there's going to be a really big opportunity to have people who have this domain knowledge in whatever the area is who can who are trained to provide feedback in a systematic way and who are being compensated via token incentives for the work that they're providing um I suspect there's going to be a lot um does this also tie into proof of humanity you know obviously like cop where the only thing that like you can sign your cryptographic keys become the proof that you there's a human behind it

so to speak or maybe yeah you're one step adding that's the third major thing I was going to get into um the the first the first iteration of rely Jeff rohf with token incentives is high mapper when most people think about high Vapor they think about getting tokens for driving um which is obviously the primary thing but a secondary thing but much less discussed although this is live today is all the map editors in uh in Hive mapper and those people are receiving honey tokens for providing rlhf to the model or in this case the map uh but they're literally going to do the edits on the map so we have the first kind of instantiation of this today uh I think there's gonna be a lot more versions of this and then the third

major area at the intersection day I'm crypto that I'm excited about is kind of this authenticity thing that you just alluded through Santiago um which is basically like there's kind of two versions of this one is hey my camera took this photo and it is unedited um this is an authentic image for my camera in order to have that you're gonna need to have a private key embedded in your in your device and I expect phones to add this in the next one or two years uh but the other version of this is going to be hey I made this image in Photoshop with my own amp skills and this is not produced by stable diffusion or mid Journey or whatever um and I need some sort of way to authenticate that as well um in any of these systems the planet key is not like that hard conceptually

up you have a private key and sign the message but you need to when anyone else uh to use that image and looks at your signature you need to know based on some sort of public source of truth that that is you um and that if you try and lie and you try and assert something that's false there needs to be some sort of public uh basically demerit system of sorts right to basically kind of like uh reduces your trust score or whatever uh I believe all of that will run out of the crypto rails I I think that third one is further out the GPU stuff is happening today the rlh stuff is just now sort of didn't happen and I think we'll pick up in a pretty new way the authentication stuff for deep fakes is a little further out but but almost certainly will happen

this is great so other than AI any other like what's your take on like some of the more Legacy stuff like D5 I remember reaching out to you while I was a terrifying he said look I just fundamentally think most value is going to accrue to the L1 so we'd rather not to like default interesting but from an investment standpoint rather touch the L1 look I'm still excited about defy uh however I think you can't put the cart before the horse here and what we've been doing as an industry is saying that people will use defy because of d5's own properties and I think that's really hard because the problem is you have to get your money on

chain first and that has a lot of friction if set up sign up for an account somewhere you have to transfer some funds in you have to buy something then you have to bother to withdraw to your own wallet and like a lot of people are just not going to go through that flow I would say we've probably Tapped Out the population set that is willing to go through that flow where they're willing to send money and get it on Jane so I think D5 is exciting once there are more wallets that are active that use defy not because of some philosophical Properties or because they actually fully understand the value prop of it most people won't understand the value prop of it but they'll use it

because it's the cheapest most convenient thing to do because otherwise they would have to go through a painful off-boarding process to get their money out of crypto and that is where the friction is so what does that mean I think what what that means is that for defy to take off we need people to get paid on chain first they need to get paid you know things like payroll is kind of the first thing that people think of but uh you know I would put in the proof physical work or d-pin uh decentralized physical infrastructure networks into this where you know you're driving around for Hive mapper and getting paid on chain for doing that

or you're helping build out the helium Network you're getting paid on chain for doing that or you know you're doing this reinforcement learning with human feedback that Kyle was talking about and you're getting paid on chain for doing that and now your money is already on chain and you might as well use an on-chain Financial Service rather than specifically how to go seek it out so I think D5 is exciting but it's like second Horizon right I think we need to focus on right now what is going to get more users paid on chain and once that hits a critical mass I think that's when D5 starts to get much more interesting

do you bucket um gaming and are you excited about that being also a catalyst for people to earn yes absolutely and then that being the primary onboarding mechanism absolutely I think gaming can be can be really big um now you know we haven't invested in a lot of games but that's yeah not because we don't think games will work it's because we doubt our ability to underwrite how fun a game is going to be before seeing the game uh and so you know we we don't like to spray and pray we are high conviction investors and if we can't build a conviction that this game is going to be fun like how can you underwrite it it

doesn't matter how good the token design is or you know the fact that items are nfts or something if the game's not fun the game's not fun that's not fun it's not gonna work yeah walk us quickly what you're like one minute interpretation or how excited you are about bass this is coinbase L2 for anyone sound familiar yeah it's an interesting experiment um I don't know how exactly uh they're thinking about the regulatory aspect here with a centralized sequencer this has been you know one of my big open questions about layer twos uh right if it's a centralized sequencer and not even a permissionless

centralized sequencer where it's like whoever Stakes the most it's just there's this company and they are the centralized sequencer well then is this basically from a regulatory perspective a centralized exchange with extra steps uh right and so curious to see how that interpretation plays out uh you know I I don't know what will happen I'm excited that coinbase is trying something Innovative and trying to do something new uh but the regulatory question here is pretty thorny I think that applies to more than just bass I think that applies to the other centralized sequencer l2s as well so would you feel differently about it Tushar if they did add if they did start to decentralize the sequencer or like

what is the what's the what's the real hang up there is it the business model of l2s or is it about decentralizing the sequencer or a little bit of both uh so I'll focus on decentralizing the sequencer uh and and I'm glad you asked that question the question is like why haven't they decentralized sequencers yet and it's because it's extraordinarily technically difficult because as soon as you have leader rotation now you have to come to consensus again and now you're back to building an L1 uh right so you face all of the same engineering challenges that you would face when building an L1 uh as soon as you start to decentralize the sequencer so you know I'm I hope that some computer

scientists who are way smarter than I am figure out a way to do this stuff I don't currently see how they're going to be able to decentralize a sequencer without running into the same engineering problems that l1's run into and then if you're just you know going to face those same engineering problems you might as well use an L1 that makes those decisions right like why are you an L2 at that point okay so like let's say so I don't know if you've talked to the espresso team well I guess like there are a bunch of teams working on like decentral helping folks decentralize the sequencer let's say people figure out how to decentralize the sequencer what are your thoughts just on l2s in general and I think I'm I forget if it was Kyle or Tushar I think one of you guys wrote

something about like just the broader business model of l2's lack of like value accrual lack of like fees being generated what I mean what are your broader thoughts on l2s uh yeah lots of us here a lot lots of thoughts here um so uh people interpret a lot of my commentaries being super negative l2s and that's not quite right it's more negative EDM than it is outfuse it just so happens that almost all l2s are EVMS uh but like the the kind of core problem is like and I've been saying this for years is it like lack of parallel execution

um yeah you're gonna have a bigger single computer on an L2 than you can for a Solana validator like sure instead of sending three thousand dollars a month on a server you can send 25 000 a month on server but like there's a limit to how fast you can execute transaction sequentially um BVM just isn't designed to be parallel execution and like if you look at where all most of the games have come from in Silicon in the last 15 years the substantial majority of being come for parallelism and so I just couldn't see if fundamental mismatch between like the core execution strategy yeah and like what actually is physics and silicon

um and so a lot of our our reason we have met bear if l2s publicly is because it's presented in the context of a way to see all the EDM but it's really just the EDM like it's not actually anything new there are exceptions things like fuel are intellectually interesting um but most part I'm gonna do with like by l2s they're they're not referring to see all the referring to evm l2s um and I don't think any of those l2s are actually addressing the scaling problem from first principles uh whether you have an optimistic roll-up or a ZK roll up you're not actually making the execution of the transactions faster um you are in the absolute most optimistic case compressing some of the data that ends up on the L1 which can reduce your gas cost for transaction but

reducing your gas cost per transaction does not mean necessarily increase TPS uh you still need to execute all of the transactions um we actually already have the counter factual for what happens when you get a big server and you turn up the gas limit on the evm that's called BNB chain um like we have the real world case study for this and no one at binance even suggests they can break 500 TPS right um and this has been like what they've been iterating on for about three four years now whatever it's been um and so I just don't really understand this this kind of book premise and I don't think it's going to get you to where where you want to be um so I think that's coming hey comment B is a little is now not a comment on

you yeah uh I just think that if you are in an ecosystem where l2s are the predominant failing strategy I believe that means l1s are overvaluing l2s are undervalued um because at that point all of the Mev um is going to the L2 and that means all of the basically that Dynamic Gap execution and also going to all dl2 um and you're basically reducing all the volatility of commodity Vans to the L1 token and uh that's bad for the value of the L1 um and so you know like if you believe I'm wrong about PVF evml2s are going to be there's going to be 10 000 adml twos and the future and that's going to be

the future of scaling um if that's what you really believe then like I think the right play is short knee long all the altitudes um because they're going to have there's like yeah basically correct it's another deal worth of zero value to be clear but it just means like you're taking all of the excess profits that are currently being earned by the online the l2s are are capturing those profits basically um and given like the size of the current valuation discrepancy between now two's and l1's is like roughly on the order of twenty to one um like your margin for an error in making that assessment is so large it just seems like a phenomenal relative

value to it uh Kyle I wanna on this point uh you one of your tweets is fairly controversial uh which said that you didn't prioritize it already and it wasn't well well I mean if I were to single-handedly think of like one of the more referral good point one of the more like um controversial I think was when you tweeted hey look security is not as important um and maybe it was taken out of context let's I want to dig into that as you think about kind of how people build infrastructure and how you prioritize kind of you know speed and other criteria over security

um so so let's dig into that for a moment because it ties in I think into all this L2 discussion uh sure some 3D referring here is probably two years old at this point uh it's in the gallon dog contact the Thor chain which uh had had a number of packs above threesomes or something I can't even remember the house but it's been at least a year and a half maybe two years um I slowly stand by the court ethos of what I said uh which was basically that that security is more important excuse me speed more important security uh there's a couple of caveats around that uh the most important of which is that's not a universally absolutely true statement for all saying he's at all points in time

um the world I live in is predominantly early stage experimental high-risk things um and I can make the argument that basically everything in crypto is still under that moniker in some form or fashion uh and and so my view is like look like you need to build a thing and get three go move fast and break things basically uh there is a point in which I've already given asset or any of the technology that that stops being true and you need to move to a more conservative methodology around development security practices uh where exactly that point in time is is pretty hard to ask for Shane ahead of time and certainly varies depending on the nature of the technology

social variables um so I don't want to make any like generalized comments across types of protocols with Technologies uh so don't stop being true at some point in time uh and that pointed out a different two different assets um but in general like I think people are too afraid of us bugs yeah no I don't mean like look some people get hurt when there's bugs and like that's unfortunate but like due to wind you got to move that and uh I think people dramatically under appreciate the importance of speed I want to tie a couple of things together here on the security comment as well as the L2 uh conversation uh I've asked a lot of people you know

why they want to use l2s and they say they want to use l2s because they say they believe that the L2 inherits the security of the L1 but this sounds very abstract and theoretical right especially with a centralized sequencer because how do you actually inherit the security of the L1 um yes things can't get reorbed once the L1 has confirmed it but that's not really the Practical thing that that most people are worried about what most people are worried about is censorship what they're worried about is that my transaction will get blocked or they're worried that a someone will have preferential access to you know front

run your transaction and that's the kind of security that they're really thinking about uh so you know I think that uh this perception that an L2 inherits the security of the L1 is too broad of a perception uh and yes it does inherit some types of security primarily on a double spend attack but it does not inherit the censorship resistance of the L1 and I think that's a major kind of flaw in the marketing of inheriting security on that is like people say you can you can withdraw from BL2 whenever you want be reflective of how malicious the you

know the sequencers or whatever and that's true but then you're ignoring the reality of like well how long does that take and like what's the opportunity to ask of your time and like if you are wanting to make a move and you can't and you're like oh you have to wait four hours or you have to wait a week like there's a good chance not four hours of that week is really really really costly to you um that could be worth five percent it could be worth 70 too uh and like it's just like explicitly ignoring the nurse enough time or priority um already notion of uh time sensitivity in these things and I like totally just object I promise if we were to um put Solana here as well I mean

there's been a number of restarts in the network um to the point where if you're a D5 protocol in Solana and this happens I mean it could have serious implications in terms of you know being liquidated the restart of the network can be problematic I'll push back too short for me in that case I get your point around like security can mean many different things for different people but the real work to me is perhaps the most fundamental of why you want to build anything in the first place in crypto like you know I guess this you know if there is a risk of a reorg which all these have not just in you know across ecosystems I mean I think everything is is you know there's still a lot of infrastructure to be built but I would think that the reords is if you can't

solve that then why even building a crypto in the first place oh no no I'm not saying that you can't solve that actually solving reorgs is one of the easier things right like you have faster finality and proof of stake systems uh and you see that with basically every proof of stake system out there whether it's Solana or algorand or uh Cosmos et cetera right like they finalize every block and now that block is final right so that solves a lot of the double spend problem for me most of the value prop of crypto is not necessarily that it is censorship resistance it's permissionless access

it's at anyone can show up and start using it without having to talk to some centralized sequencer and hope that they're treated fairly uh it's also that anyone can start validating the chain without hoping that you know oh I I hope I I'm allowed to um right yeah on that point I mean there's two things that come to mind one is how easy is it for someone just send off a valid a validator right and obviously Solana has the hardware requirements it's been kind of a criticism right the hardware requirements are more it's more expensive it's more difficult than in ethereum but um so that's one the other one is just on a nation state level kind of in a regulatory front you know what are your thoughts around censorship at that level where it can be

problematic if you're in a particular jurisdiction I mean China's always been the place where it's difficult but as the regulatory environment evolves it could be difficult in in places even like in the US so how do you think about censorship in that context I think the comparison of eth L1 to Solana L1 validator costs is not necessarily the best comparison because the vision for eth is to have the transactions going through l2s so the question is how much does it cost to spin up a L2 validator node for one of these Roll-Ups oh wait it's actually infinite you're not allowed you are you are not permitted to do this and if you

were it costs more I would bet I don't have the numbers in front of me because this is all private company information I don't know how much these centralized sequencers spend on their server costs they don't reveal this stuff but I would bet with pretty high conviction that it costs more than running a salon of elevator um so I I just don't know that I that I buy that comparison hmm too short and Carl do you guys have allocation to eth just like eat the asset not not the ecosystem but just eth we do yes what's the percentage breakdown of like eat to Soul uh we don't talk about relative position sizing and actively traded things

publicly fair enough we're consider there's more Soul than there is Eve uh but but we're not gonna get it we're not going to get into whether it's two to one or ten to one or whatever fair fair okay um all right so I'm looking through some of these uh topics that Jr put together and just like some things that might be interesting to talk about and one of them that I found pretty surprising was [Music] um I don't know if one of you guys wrote this or Jr wrote this but filecoin emerges as a meaningful ecosystem on par with eth and Solana in terms of legitimacy I would say that would uh is thrown in the Bold take uh department so I yeah I definitely need one of you guys to expand on that uh I guess I wanted this one uh

I'd hated final coins for a really long time uh and I've done a 180 in the last six months uh my my the primary reason and I mean these things coming together um one is after blockchain launches my general obligation is going to take one to two years sub blockchain to like become legitimate and do do what it's going to do and conversely if it doesn't become legitimate over the course of that one to two years it typically falls into irrelevance and kind of dies um uh it doesn't necessarily like mean it stops producing blocks but it kind of like falls out of public purview and Consciousness uh filecoin is two and a half years old now so it kind of just

passed that that threshold and so I use the opportunity in September October November last year to start kind of diving into the public when ecosystem at the two-year Mark uh and was was pretty pleasantly surprised that like it was very clear that the the rate of intellectual Capital contributing was increasing um so that's kind of an a comment B is uh without my virtual machine launch and this is probably the most important thing uh the falcoin virtual machine launched uh a month ago on Pi Day so March 14th and um development virtual machine is the EDM on filecoin um that does not mean like I mean maybe there's going to be D5 point I don't know that's going to call that thing or not that's not the right interpretation

of it though uh the right interpretation of it is you will be able to uh pre-program payments using UVM and also see what other other VMS as well hopefully if you ask me um uh and those will be able to connect to file for Native bindings for the kind of three major Primitives in filecoin which are storage uh retrieval and bandwidth and compute uh and uh you'll be able to reprogram those image which means you can do things like carbon storage that means you can have payments for retrieval markets that means you can do uh additional payments for specific types of compute workloads um but basically in like it makes filecoin programmable in the same way that like Bitcoin wasn't programmable and then each made Bitcoin programmable

not literally the Bitcoin asset but like gave you a programmable asset Ledger um we are now introducing the notion of uh programmable payments onto generalized compute operations for kind of the three basic Primitives and computation uh and that's a really big mission uh and so the fbm is kind of kind of thing number two um and then uh I think number three is retriever markets went live um file points uh Native uh retrieval market is called Saturn and launched in January of this year very quietly I think there's roughly a thousand nodes on it right now uh today Caymans are not enabled on Saturn so for now it's all still in kind of like play muddy test mode it's real data but play money

um but obviously at some point they're gonna turn on payments and get this thing going and uh I'm very excited for the vision of a d75 CDF it's been like one of the sexiest kind of uh North Star visions of crypto for a very long time uh and we're finally getting closer to that I think as that comes closer to becoming reality we're gonna see a lot of really amazing things both on top of it so that kind of culmination of things has gotten me very excited about filecoin and like those are like all very direct file coin comments one question on that is like uh I remember the 2019 multi-coin event um multi-coin Summit in New York the uh I forget his name the are we founder I don't know yeah red killed me on are we've got like pretty excited about our

weave um and it seems like some cool stuff has been happening with our weave recently um how do you think about like file I mean they're both decentralized storage networks but my understanding is like filecoin has proof of replication and proof of space time are we've has like proof of access I think they call it they have different economic models um maybe different Target use cases like filecoin seems more like almost B2B like large amounts of data like archives backups are always more like social media posts and like web pages which is maybe a little more b2c like how do you think about the difference between between our weave and filecoin uh

the primary reason we gotta tell them are we back in 2019 was it Simplicity um at the time five point literally didn't exist um and we understood even back then that like Falcon was going to be a really complex animal when it did launch uh what we liked about our weave was pay once throw it in the sky it gets stored forever uh and you don't really have to think any more than that and there was there was a real value to that Simplicity uh and that assessment was fairly uh not popular at the time although I think a lot of people came around to appreciate the importance of that um uh over that kind of called 24 months um A lot's happened though that's been

four years since we yeah it's been four years since we won our first R weave token so I think we left them in May of 2019 uh those in a while um and in that time the product ecosystem has matured it's been a lot more usable um today I look at these things uh today I gotta look at filecoin is basically being a superset of our weave um using the fvm you can build a permanent storage system that is functionally you know offering the same type of guarantees as our weave what's nice about the fbm is uh it is you can parameterize it in ways that you can't parameterize or a week or we've isn't really parametrizable um there's one parameter which is paid

for a file and then however many people replicated because of like the crypto economics of the r weave proof of access gain is how many replicas there are and there's no notion of measurement or concreteness it's all probabilistic uh as is one way to parameterize permanent storage uh filecoin gives you infinite more optionality of the fdm where you can say hey I want five copies of this I want 25 copies of this I want three copies and use confidence um you can parametrize how you want it to be stored in any way and because you have groups of replication uh Griffith's face time you can enforce the different parameters um in the system so I want to have that optionality creates more complexity in terms of like

you have more decisions to make as a developer on the other hand oh saying there's one permanent Source configuration for everyone in the world actually is not not right to me I think you need to be able to parametrize these things and so um I look at them as they will coexist for a long time but the file coin definitely is vastly more configurable that our wave is yeah guys last topic I want to touch on is is Bitcoin you talk about stuff that you've changed your mind on this is something that I was I've changed my mind on recently and paid more attention we have minibon talking about you know changes pretty fundamental changes in the Bitcoin ethos at least to explore you know Define Bitcoin and other kind

of um proposition so I'm curious if you guys internally have paid attention to that if that's something that you know you get excited about you're investing in the Bitcoin ecosystem or just increase your allocation of Bitcoin we do own some Bitcoin uh but you know I would say that the benefits that Bitcoin has are potentially shared by other L1 assets as they become more mature the key thing that Bitcoin has is it has a big head start

it's been around for a while right and I remember in 2020 I was talking to you know a bunch of traditional investors you know the the Wall Street types and the perception then was that yes traditional investors will come into crypto but it's gonna be Bitcoin only they're really only looking at Bitcoin they like the digital growth gold thesis because they understand that and you know that that's really it then something really big happened in 2021 which is we saw that those big institutional allocators said Bitcoin and eth and now everywhere that you go look at Bitcoin it's kind of Bitcoin and if

uh basically in any portfolio any Institutional Investor right like eth has caught up with Bitcoin in a meaningful sense in terms of liquidity in terms of recognition uh and in terms of a perception of neutrality yes it's not as you know neutral as Bitcoin is metallic didn't disappear like Satoshi did they still do hard forks and stuff but the perception of it is Bitcoin and eth on similar footing eth is what five six years behind Bitcoin so when you play that out right like right now you can say Bitcoin has been around twice as long as Eve a decade from now it's you can't say

that anymore right Bitcoin was around five years longer but you know both things been around you know 15 years in in 20 years right so I think things change over time and you know now I'll expand that I don't think eth is the last one to get into that club of Bitcoin and eth I think that the most used smart contract platform will get into that club and it will become neutral over time uh or perceived as more neutral especially as hard forks slow down or stop as the tech ossifies uh and it'll have more users which ultimately I think is the thing that matters the most is having more

users getting more people paid on chain like I was talking about earlier you're not going to see that on bitcoin you're not going to see more and more people get paid Unchained Bitcoin that's extremely unlikely um but you will see that hopefully on Solana or hopefully on ethereum or you know another smart contract platform that does more useful stuff so yeah I think Bitcoin has its moment uh right now with uh you know inflation looking like it's sticky and a whole bunch of macro things that I do not consider myself an expert on um but I think that it's not as uniquely positioned to capture those Tailwinds as

some might like to believe yeah fair enough I want to end this uh just with a question that um I had been my former partner verify here last week and I asked him this question and I'd like for you guys to also answer this is what would you tell a younger version of Kyle and Tushar entering the space today did you wish you knew kind of in however so many years you've been in the space follow your heart man like don't don't get caught up in uh the drama and the pressure of crypto sweater it's funny you say that out because Ben I'm sure he has an ALT account but he doesn't have a Twitter profile I would say

your job is to earn excess returns and the only way to do that is be contrarian and be right and that requires bravery like you need to be willing to be wrong in order to be able to be contrarian and right so don't be afraid to be wrong in public well I think that's a good way to end it um guys thanks for coming on I think it's been a while and it's great to hear from you guys um so appreciate the takes and and you have guys answering some harder questions uh um we'll certainly have to have you on later on in the year as things progress and uh but otherwise really appreciate the time and and uh yeah thanks for coming on thanks guys appreciate you

having a thought guys [Applause] [Music]