The New Rules of Crypto Distribution | Roundup

Transcript

The reason why you want to focus on distribution. So again, not enough founders I speak to understand this or think about this, but if you try and piggyback off someone else's distribution where they have the power over the economics, uh they're going to compress your margins and they're going to have the final say around your economics. If you build something and then someone has users like Coinbase or Binance and you say, "Oh, I can build this for Coinbase." And they can share it with their users as distribution. That won't work because they're going to take all the margins to themselves. they're going to squeeze you and probably the product is highly undifferiated and so it becomes a commodity and then you end up trending towards zero. And so I think again here this is why it's so important to build distribution yourself is because when

you own the distribution you own the end user you own the economics especially in the application space which is where we're at right now in crypto. This episode is sponsored by Canton the only public permissionless blockchain built for institutional finance. We'll hear more from Canton later in the episode. Blockworks Digital Asset Summit is back in New York March 24th through 26th. We'll have top speakers from leading asset managers, financial institutions, DeFi protocols, crypto companies, and policy makers all under one roof. Think Black Rockck, Coinbase, Robin Hood, and more. Follow the link in today's show notes to buy tickets, and make sure to use code Bell200 for $200 off at checkout. See you there. Hey everyone, quick disclaimer before we get into today's episode. Nothing said on bell curve is a recommendation to buy or sell

securities or tokens. This podcast is forformational purposes only and the views expressed by anyone on the show are solely our opinions, not financial advice. Our guest and I may hold positions in the companies, funds or projects discussed. All right, everyone. Welcome back to another bell curve. You got me, Zave Miles. Miles holding it down from snowy snowy Denver. >> No, it's sunny. It's sunny Denver. I show up here like with all my Boston winter clothes and it's like 60 degrees um you know in the daytime. But um yeah, no complaints actually on that on the weather side. Um and say it's uh it's been a good week. Um

>> what are the vibes like there, Miles? >> It is it is quieter. Like I've been coming to Denver for um I would say this is like my third or fourth year in a row. Um, and most prior years I've never even stepped foot in the main event. Um, because there are so many people and more like interesting side events going on um that you can just meet with much easier like outside of that, right? This is the first year I've actually like there hasn't been like enough people or side events like for that to kind of be my my um mode of operation. And so yeah, I was there today and I would say like um maybe I'm just getting a little old

but um walking around the booths like noticed that there are less names that are like widely recognizable like big DeFi names um that you would expect to see at these things. um for the most part there's not a lot of you know talks or activity outside of the stuff that's already working. Um so I can probably classify that as like um stable coins like everything crossber you know basically the fintech intersection um rwas in defi um and then I think outside of that it's all the you know the I

guess the the one frontier that people are super excited about which we've covered before in the show which is like the intersection >> agents paying with stable coin points. Um, and everybody having their own, you know, flavor of every need that is needs to be addressed in that sort of, you know, the entire stack, whether that's like, you know, layer ones that just, you know, have a bunch of judges to determine whether an agent did the right thing or not, or like reputation systems for agents. Um I would say less on like the decentralized training um than you know maybe even a year or two ago um you know obviously like X402 and then all

the protocols um that are being built uh to like set the standards like that's that's very you know present and and being talked about a lot. Um so yeah it's uh it's it's a weird one. Um, I'd say better than conferences where you come in like the thick of frothy bull market and it's just there's so much noise that it's really hard to filter through like you know there's definitely like a better signal to noise you know ratio today than there was probably last year. Um, but it's um I wouldn't say as like the the level of excitement is nearly where it was. Um, I think like two years ago was actually the best ETH Denver, but um that was like in the

middle of the the infra uh before all those infra tokens got crushed. I um I mean Zave and I were at Eat Denver last year and had this fateful talk that I keep giving Z credit for publicly, but uh it was also I mean to me that feels more on the money than what people were interested in couple years last year which was I mean there was this AI crypto intersection but it was very nebulous and it was decentralized training and inference and bots and all of this stuff and I would I would say I'm still very bullish on AI creators. I I think AIXBT, you know, not necessarily that ticker or project, but things like that are super

interesting to me. Um, but people are really interested in Bitcoin L2s, ultimately. I mean, I've said this before, I think we're still a little ways out from onchain agentic payments in crypto, but it's obviously going to happen and it'll be big. Someone will make a ton of money facilitating that. Um, and RWA are very real and it's kind of the thing that's working. So, that makes a lot of sense. Miles, give us a sentiment pulse check. Like one is fullscale crypto winter bear market, no one's happy. 10 is, you know, people are people are renting yachts for the for the after parties and stuff like that.

Where where are we on a sentiment scale? >> The people who are working on the things that are working like are actually be like showing, you know, we'll talk about this later, but like where there's market pull, right? like the stable coin issuers are inundated with demand, you know, from institutions, from chains, from everything, right? Um they seem very happy and but it's less like the usual crowd that's happy when things are working because those are of course like you know companies, equities. There's no tokens. There's not like people are like speculating on on this you know uh use case that's you know kind of a um exploding and and very successful. Um,

and then I would say like there's uh a sentiment of um, you know, optimistic sort of and and hope for like the AI side, but you can tell it's not like, yeah, guys, like look at the charts, they're going up and to the right or look at the amount of demand we're getting for this stuff. Um, I think that's the long and short of it. Like, you know, the the it's it's it's less of a speculative like positive sentiment. and it's more enterpriseheavy uh for the stuff that is working. Um and you see very I saw like nothing in terms of like social fi or like you know it's kind of reflects like the crypto Twitter to some extent. Um like outside of financial use cases. There's just not a lot of stuff

here. Um now there is still some weirdness, some crypto weirdness which I love. I saw people >> in some very interesting outfits. um saw, you know, one of the booths where you step in and they have like dollar bills flying everywhere and you try to like grab the dollar bills. I'm like, okay. So, like the mercenary capitalist still has not totally left the space. Um and the dens have not totally left, but like yeah, there's definitely less of that. >> Nice. um >> quiet from the Cosmos crew who I've always loved I've always loved spending time with them at conferences because they're always kind of like yeah like we're we're ahead of the curve and now they're just gone. It's co >> the the Cosmos crew has moved on.

They've already won, mate. There's no need for them to spread the good word anymore. They've already won. They're sitting back relaxing. They've they've seen how it all played out. [laughter] And >> um >> one is a pretty interesting way to describe what happened. Uh, as a as a Cosmo sympathizer, I feel like I can get away with saying that, but all three of us are a big Yeah, I love the spirit. Exactly. >> Um, yeah, I I think it's interesting insight, Miles. I I think one question to ask, I guess I'm not sure if you were at Dev Connect, but uh, funnily enough, the most recent Ethereum event was in November last year, and there was actually pretty few cryptoai like side events and meetups, which surprised me.

Like I went to the event thinking there'd be a lot of this and there wasn't. Maybe in the last few weeks like post open claw I imagine there's a lot more sort of craze and hype and you know more like stripes etc entering the space bigger than ever. >> Is are there a lot of events on cryptoni now or is it like still small like it was okay I mean that's been a big turnaround then because a few months ago was very different. >> Yeah. when I looked at the side events or I looked at like the dedicated stages like um that it's dominating um and it's and it's but it's different right it's not like as much about decentralized training or like you know the the originals probably what you know was taking up conference space like two years ago um it is more like talking

about how do we solve the trust problem for agents we we you know these are great payment rails for agents but everything around it like how do we manifest this future that we think, you know, represents the perfect marriage for crypto or like, you know, solves this identity crisis for crypto and what it's useful for. Um, and that to me is the closest thing I've felt like to, you know, the sort of attitude or like, you know, in atmosphere of like two years ago when it was like, okay, you know, reststaking opens up a new design space. Uh, co-processors open up a new design space and like you didn't know just exactly how soon or if this would be useful. Um, it would, but but it, you know, that's where people were clearly

like placing their attention bets. I have um I'd love to get into maybe that's a good leadin into this week's topic which is distribution and we're going to tackle it from a number of number of angles. So uh maybe the strategy from an infrastructure perspective from an app perspective and then I think AI and what does distribution look like in the world of AI is a really interesting topic. I I would like actually if we can maybe get 10 minutes at the end actually one story I'd love to get your guys take on just because we had this whole debate about hyperlquid and buybacks and everything.

Did you see the hyperlquid policy [snorts] center that got set up with seated with $28 million that Jake Travinsky is leading? Uh I have to say I feel like this is an incredible use of funds. Um this makes an enormous amount of sense to me. We were just talking about on the last episode so I want to bring it up. Yeah. So Jake was previously um at Variant and he announced a couple weeks ago that he left and then that he'd be leading this policy institute and he was seated with a yeah uh $28 million worth of hype which >> super interesting. Okay, this is not going to buy back hype. It's actually going to fund um you know the regul regulation side for them. Okay,

>> I didn't I totally missed that. >> Yeah, >> that's big. That's that's high signal. Jake is amazing. uh Variant is an investor in Delta and he was our go-to person for anything >> uh regulatory. Um he knows everybody, trust him. So that's a huge win for Hyperlquid. >> He's a beast. It's him and Rebecca. Um I feel like is um I'm sure there are there are many more, but those are just ones I would have a high opinion of. But yeah, it's I think it's a massive win for Hyperlid. There's a great example, by the way, of I feel like I would um as a hype shareholder far prefer 20 million to go towards this than a buyback where I feel like the upside here is you know many multiples of the 20 million that

you had spent, right? Um I mean speaking of I guess we mentioned you know how you could use the cash instead of buying back the token last week. uh Jake, you know, a lot of legal experience um in terms of like how they're redistributing the cash. I assume the growth now they're looking at is actually that regulatory front and that's what we were chatting about last week as well where it makes sense to use that cash to lure the best people in there and to like help them outgrow you know just a kind of cryptonative audience but actually push into larger global markets that are regulated and actually start competing with Binance. So for me it's actually no surprise whatsoever that they're using the cash like this to yeah grow their market share like that.

>> It's a good use of money and I mean from at least my perspective and I mean this gets us you know talking about branching out and extending into new audiences. Let's talk about distribution. And distribution is a word that people have focused on for a while in crypto. I'd argue not as much as they maybe should have in the past. Um I'm curious actually if you guys look back at all at the fat protocol thesis um and with either rosecolored glasses different air in crypto has it been totally repudiated I I don't know but you know there's this kind of saying that often gets quoted of first time founders think about product second time founders think about distribution and maybe because of this I just to set up the relationship for

distribution in in crypto and relationship to the customer I think we had this we had this idea of a fat protocol thesis uh which Joel Mangro partnered to Chris Berninsky at placeholder released which is the idea that in the in the internet uh era there were these protocols uh they were very thin but they were open to everyone and they they created a lot of value but they didn't capture it and you had fat applications that had access to the customer built on top and he argued that in crypto is going to be opposite you're going to have fat protocols with thin applications eight years later it would seem as though that's not necessarily the case although you could say scoreboard look at Bitcoin Um, but it seems now that the attention has flipped and that we are 100% focused on distribution and and a connection to the

customer and we can talk about what exactly that means for different categories, but I guess just to kick it off, do you guys think differently about the fat protocol thesis? Do you think it's been repudiated, backed up in some senses? Do you think that the dynamic in web, you know, crypto is going to be the same as it was in web two? How do how do you how would you score or rate the mental frameworks that this industry has had on this subject for the last almost decade? >> I mean, yeah, I can start maybe um I think the the fat protocol thesis maybe where it was off um was probably in two places. one. Um I think it assumed that the uh early winning layer ones um would

be would have more defense defendable moes than it turns out that they they actually have in reality right um I don't think this idea that chains would could be or will be like commoditized you know was an ass was like an assumption in the fat protocol thesis which I think has actually turned out to be more true than not, right? Um, and the, you know, it basically positions like apps as this thin layer on this, you know, uh, on these networks that, um, you know, there's little choice as to which network you would deploy these thin apps to because of the network effects would be so powerful with the early winners and the switching costs

would be, you know, higher than they actually turned out to be. Um, I think that's wrong, right? And so now I think like the fat protocol thesis is you know it's it's um it's lost also because maybe developers like just want more control right and they and I think at the time uh I don't think it was uh it was underappreciated just how like uh little these networks could cater to like real developer needs right so now we have proliferation of these networks um and sometimes the apps are just you know vertically integrating the the actual chain itself Um but yeah, I'll pause there Z curious to hear things your thoughts.

>> Yeah, I think great points Miles. Um just briefly for me the some elements of the that protocol thesis still stand. Um you're seeing some of the biggest winners still today uh Ethereum, Salana, etc. Uh I don't think these are going away anytime soon and they have captured a lot of the value of the applications that have launched on their chain in the first place. I think Miles you mentioned something you know really important there around the kind of modes of these fatter protocols let's say like Ethereum and Salana Ethereum you know the way it executes and runs the compute is is simply EVM and I think they came out first built a really solid developer ecosystem and then they got a lot of liquidity and that grew that outgrew all

competitors essentially and I think you haven't really seen another EVM chain yet reach that same velocity the reason why I think Salana has staying power is again because it kind of flipped the EVM on its head and then created the SVM and then they had a lot of developer network effects super fast therefore liquidity more apps more users etc and since then you've seen again it's been too hard for other SVM projects to compete and so what happens is a lot of the best applications flock to these winners and so they capture a lot of the value from having those winners and saying that I'm also of the impression that we're in the very early innings of the application era and so I do expect that we could have winners potentially as big as some of these fatter protocols like Ethereum and Salana, but we're in the very very

early innings of this which we'll explore in this episode around like what are these apps, what have they done well, like how important is distribution and can they actually capture more value than the the protocol underneath itself? And I'm of the impression they can long term, but I think we're still in that transition period. Before we get back into the episode, here's a quick word from our sponsor, Canton, the only public permissionless blockchain built for institutional finance. >> Tradfi wasn't built for 24/7 markets. [music] DeFi wasn't built for regulated institutions. But Canton is where the power of finance meets the promise of blockchain.

The only public blockchain [music] with institutional-grade privacy, true composability, real assets, real markets. Why reinvent finance when [music] Canton makes it work the way it should? Canton, where finance flows. So I think we've talked about this before, Miles. You and I've done a whole episode on this just aggregation theory and the smiling curve that people have pointed out as a way of capturing value across um or as an ind as as a rough way of quantifying a relationship of value capture across supply chains. But the idea just being if you were to look at

the beginning and the end of the supply chain, those typically do well. There's lock in. Um, this was originally used to describe the computer industry where there's a lot of patents and R&D competency that gets built up and then the end where you have the access to the user. It's the middle that gets squished. And that's been my at least theory with crypto for a long time. And one thing one thing though that I've been continuously sort of surprised by and it's in my two hard bucket now of I mean just thinking about assets like ETH and Soul. Um they they have I think definitively you know people have always said it doesn't matter you're not going to be an Ethereum or a Salana user and I think that just hasn't really turned out

to be right. Um I think that those ecosystems are kind of more siloed than you might think. there's a cultural element of people thinking of themselves as an ETH or a Salana person or something like that and a lot of the like you kind of see something like Phantom which and MetaMask which they're now interoperable with their other networks but it's kind of like too late and yeah I don't know I think um uh I we've talked about on this show I think that it's very hard to understand really how to think about ET solely assets at this point because the barriers to entries are so high um for new competitors which feels good but yeah on the distribution side um I think it's going to consistently be some of

these new applications that end up yeah having um you know they'll that leverage of having direct relationship with the user is super super important um why why don't we talk why don't we begin on the app side of things I'm curious this is kind of the big question that everyone has at the moment which is how do you gain distribution um and say if you work with a lot especially new application teams. So obviously if you're Coinbase or Binance, great. You have distribution and you leverage that uh from a product standpoint, but let's say you're starting out today, right? Like maybe I guess you could give a rundown of what has historically worked on the distribution front and then how are new apps thinking about it and then Miles, we can talk about infrastructure a

little bit. >> Yeah, absolutely. Thanks, Mike. Um, yeah, I think right now in terms of where I spend most my time, there's kind of two fields. one is the crypto and AI intersection and the second is the more consumer app side where I think there's a lot of opportunity right now in crypto to build really fantastic outcomes there. So on the application front um you know we have invested in many apps already at practice and whenever a team comes out now to pitch us I was saying earlier in this call that you know with AI now it's never been easier to build uh any product almost and ship that faster than ever. So there's a real question of like the market itself and what the market wants and not just what the market wants, but how do you get to the market in the first place? That

really for me is the key question that not enough startups ask themselves before they go out and come up with an idea for a product. Um, and if you look at all the winners in crypto, they've actually all had unique distribution strategies. And I feel quite strongly about this point, the fact that like to win as an application, you need to have a unique novel distribution strategy that hasn't really been tried before. Or maybe it has been tried before and you have a new spin on it. I don't think you can copy someone else's distribution strategy because I don't think that will work differently to infrastructure which Miles will talk about in a minute around that is a bit of a playbook that people try and kind of uh emulate from each other on the application side. I think all the winners have done it differently

and so I'm just going to go through a few names here which I've thought about a bit you know in the past few months of what has worked for them and then as I work with portfolio companies trying to innovate around maybe taking some inspiration from them around what has worked and what hasn't. So with hyperlquid I think we can start there as you know a team that obviously is successful right now on the application side here is quite interesting because they borrowed I think some of the infrastructure playbook to build their own community and distribution. They built a fantastic app and then you know air dropped a lot of tokens to their stakeholders and made them quite rich and then that made them talk a lot on social media that really like pumped them up. Uh that gave them their own sort of native distribution from the

fact that so many people had a stake in that game that they wanted to share about why it was so good. Then people on boarded the product. Mike [laughter] >> I just I'm curious because this is it's so hard to do the counterfactual of this. Do you think that the airdrop was critical to their distribution and success? >> 100%. I think if Hyperliquid wasn't okay, it's all related, but if Hyperliquid wasn't a great product and people didn't get rich, I don't think it would have had as much distribution. And I think they've kind of taken that playbook a bit from Ethereum and also Salana in terms of early holders getting quite rich with also really good technology that people can get behind and clearly like enjoy using. So I think Hyperlid is the most similar playbook I've seen on the application side to like historic playbooks that has worked.

I think they're going to be the only one that works here. I think people are going to try and copy this playbook now going forward and it's not going to work. So congratulations to Hyperlquid for getting on that early because like I could be wrong but I really can't see another distribution strategy similar to Hyperlid working like that in future. The second one I want to talk about is is Poly Market. Poly market, you know, when they started building out, they got really good distribution through media uh because people were saying that their stats and facts were better than the polls and there was a narrative around that and so people Poly Market stats and odds were popping up in media places and so they were getting a lot of eyeballs even from a non-crypto audience and Poly Market is one of the only teams in

crypto that has gone really non-crypto and the way they've done that is through traditional media which a lot of crypto cannot crack into. So again a very novel distribution strategy which no one had attempted before poly market really went live and so again very interesting to look at. Athena another successful app I think they've done an incredible job especially with cryptonative distribution to start with and then in future like more distribution towards banks. They built a very strong brand which is trusted in the early days which I think is a good start to build any distribution strategy. They align stakeholders very early a lot of Ethereum Wales etc. people very native to DeFi that had a lot of capital already so they could start that flywheel and get that money in very very early. The marketing was great and they also got a lot of distribution through

exchanges. I'm thinking like Binance etc. So they align those partners very very early on. Athena's distribution strategy again here completely unique. No one has tried a distribution strategy like Athena um especially in the last year or two and again I I can't see anyone copying that strategy and being successful. The next team I want to spotlight which has raised capital from a a traditional fund more recently is called FOMO. So FOMO is very very simply a social trading app. Uh again here like a novel distribution strategy at least in crypto borrowed from more web 2 and trad sharing easy onboarding great UX commenting on other people's trades.

It's all all the things I'm mentioning right now very social and I'll get back to that in a minute after I finish naming the list. The next application that's done really well recently is Axiom. Axiom again here, completely different to everything else I've just mentioned. So, one thing that Axiom has done is cashback. Again, that's not like a massively novel distribution strategy, but at least the way that they've executed it in crypto has been really great. The second thing they've done is actually referrals, which in crypto hasn't really worked that well preaxium, but they've done a really good job of like referrals, earning from people you refer through, and I think they've incentivized their community very, very well to like spread the word. You always want users distributing your products and I think Axiom has done that like really well. The the last example I'll

name for now um which is also interesting which I think people don't talk about enough in the space is the kind of like Telegram native distribution that things like Trojan Banana Gun have really leveraged and again pretty simple the thesis here around why it's worked. I mean, it's easy to copy, easy to share, easy to use, easy to add people, like it's easy to like make viral. And so, I think if you look at those six apps, which are all working quite well, and you could name some more, I'm talking more like kind of end user apps as opposed to DeFi apps. All of them have completely different distribution strategies. And I feel very strongly that the best apps going forward will continue to have unique distribution strategies. So whenever you analyze any application, you should always be thinking like what is novel about the distribution and how

can they scour that distribution more and more over time which is also an incredibly hard thing to do. >> I have a um I have a framework for how I think you could generalize the different approaches to distribution and I think Ze those are all great examples and they all basically fit. So, you know, if I were to look, I was kind of thinking through this before the episode. Um, one, and several of the ones that you describe, I think are piggybacking on an existing network. Um, so this would be, and I think there's a timing element to this as well, um, where you can kind of use it to get escape velocity, but some of the early Facebook apps I think are like a really good example of this or, um, understanding and leveraging Google

uh, like Google Adwords, right, before when they were still very cheap. There was actually a time where you could get paid to advertise on Google, which is insane. Um, but it was a time it was a way where you can kind of borrow the flow um of another network and get to a point where you become kind of the brand name. Um, there's another one that we we didn't talk about here that I would love to get your guys take on because I think actually that uh uh wallets uh do this, which is controlling a scarce uh choke point or owning the game. Probably the best example of this of all time is um what uh Windows uh the owning the operating system. So everyone wanted to use computers. You got to borrow that trend but then everyone was dependent on the operating system and then all the

switching costs compounded to the point where no one could move off of it and they just had unbelievable uh distribution that they could pump their own firstparty applications into. when they probably created the best SAS um you know SAS uh package of all time which is the office suite right like that's probably one of the best SAS products of all time so piggybacking owning the choke point um one thing that this maybe Z fit into your description of um uh uh euphoria but like uh or FOMO FOMO sorry rather um is uh ver building verality into the product. Weirdly, you know what I thought had a really good

example of this was NFTTS. It had verality built in because you'd make it your profile picture. It's kind of the best building verality and distribution into the product wasn't enough to save NFTs, but it was a really good growth hack for a little while there. Um and then the other is uh basically aggregating um and maybe we won't get to this when it comes to LLMs, but moving one layer up um and filtering and helping um uh curate information. So Google, Spotify, uh are great, but Google's the best probably example of this. Facebook though, um things like that. So I think those are the different lenses. And when I looked at the examples that you were describing there, I many of them are like borrow the flow

um like piggybacking on social or other media channels. >> And and what's interesting about Axiom and FOMO in particular is like those are not insanely novel products, right? Like they are just more or less like trading terminals and and serving as aggregators. They win on UX and they win on these DRO hacks um that are either meeting users like where they're at today. I could say the same for Poly Market too if you're like seeing data from Poly Market on CNPC, right? Um or you know basically scaling up and manufacturing this word of mouth like virality. Um, and I think that that's interesting just because like crypto for

so long has always been trying to create net new novel things that are totally different than anything has ever existed and it's rarely kind of meeting demand where it's at today, right? Um, and that is like fine because for a long time we were kind of working on things that were like unexplored and unrealized. Um, but the reality is like today as an app like there are things that are working. Um, like as we were talking about terms of like the Denver themes. Um, and so you know I think FOMO and and um and Axiom are great examples of like you don't need to create like something that's net new like never seen by the world. Hard to explain to consumers, right? Actually the easier it is to explain to consumers

is sometimes going to give you the highest chance of success. Um, and then you just spend all your time on just manufacturing like that that viral, you know, feedback loop or or whatever it is, right? Um, so we'll see more of that, but that's also why we see less crazy ideas in crypto today, I think, as well. >> Yeah, just to build on a few points you guys mentioned. So yeah firstly I think Mike you you said it well a lot of the distribution we see in these applications they're very social inherently social and I think they also build narratives at the same time and these also spread socially for example like hyperlquid was kind of like VCs are bad and then everyone like rallied around that and it became consensus VCs

were bad and that was good for distribution as well because they align with that message same as like poly market you know news are better than you know prediction markets are better than the polls I think like that as a narrative as well, especially around elections, people rallying around that, building consensus around that, sharing that as well. Like they're all very social. Going to where the users already are is super important. You know, you want to be across like Telegram X. You also want to be at conferences, you want to be in the news where people are reading news. Uh but also at the same time, you need very fundamental product things like a good UX, a good product, uh etc. And a good way of on boarding users in the first place. I don't think distribution can work unless you have actually a sound product. But building a

sound product again is getting easier than ever. So it's more important to focus on distribution I think uh than ever. The reason why you want to focus on distribution. So again not enough founders I speak to understand this or think about this. But if you try and piggyback off someone else's distribution where they have the power over the economics, uh they're going to compress your margins and they're going to have the final say around your economics. If you build something and then someone has users like Coinbase or Binance and you say, "Oh, I can build this for Coinbase." and they can share it with their users as distribution. That won't work because they're going to take all the margins to themselves that they're going to squeeze you and probably the product is highly undifferiated and so it becomes a commodity and then you end up trending

towards zero. And so I think again here this is why it's so important to build distribution yourself is because when you own the distribution you own the end user, you own the economics especially in the application space which is where we're at right now in crypto. And that's why I think people who are pushing the boundaries and the frontier of like distribution strategies, including with AI, which we'll talk about later in the podcast, I think are really going to be the winners long term in the space. >> Yeah. And I think this goes back to fat protocol thesis being right or wrong again. Um like if you notice two years ago, I think two of the leading um DeFi protocols like Uniswap and and a um you can extrapolate this more broadly. I think for like D5 protocols is different, right, than like an Axiom or

a FOMO that is literally just like the thin layer at the top. um they were optimizing for like they're like okay we have we've found PMF and users like use us directly now let's become a platform right um >> and I think we've seen a like reversal of you know those decisions and like actually you know we're the entire like we can't be this plat or like the scale of this platform and the amount of developers like in the economy that's going to run on our platform is actually not as big um or It's actually harder to like manifest that vision than it is to build a retail app, right? And get closer to the customer again and

monetize and like, you know, create a moat based off of the the the user wanting to go on their phone to the A app to the Uniswap X app, right? Um, and that's funny because as an investor a couple years ago, it was always like, you know, you were looking for the the the grandest vision, the biggest possible platform, the largest economy that could and developers that could be like, you know, launching on this thing. Um, and yeah, I think Hyperlid's in an interesting spot right now because they did start narrow and now they are trying to become a platform, but they still, you know, are are trying to balance that without with, you know, keeping staying close to the customer. But I don't know.

It's an observation. >> Everyone in crypto loves to talk about transparency, but would [music] any serious business publish its payroll for the world to see or hand competitors its contracts? That's not a feature. That's a bug. And it's why most financial institutions can't operate on today's public rails. Canon network is different. It's the only public permissionless layer 1 blockchain built with privacy at its core, sharing data only on a need to- know basis. That's why Goldman Sachs, NASDAQ, and Broadridge are already building on the network. Only Canton delivers the network effects of a public chain with the privacy and control real finance demands. Learn more about Canton by following the link in today's show notes.

You know, I was going to kind of get to this later, but I think one very tactical tip, and this is not possible for every founder in every business, but I was reading this this piece um called founders are cult leaders. Um, and generally I hate this kind of talk but sort of agree with it a little bit. Um, from the perspective of you need to unite people behind an idea. If you're doing any enterprise, if you're the head of a state or a business, you need to really unite people and motivate around ideas. And one thing that makes your distribution much easier actually, you

know, this idea of it's it's not you should pick a hard you should pick something hard. Um, and sometimes counterintuitively it's easier to do something hard than to do something easy. And I think the reason for that is because when you pick a harder, more interesting problem, you can attract better talent. The other >> there's less competition. >> There's less competition. The other reason is because it's more interesting to talk about. So, you know who I've seen do this unbelievably well? I think this guy's the goat at almost like is Elon Musk.

Elon Musk's always talked you got a he's a reality warper, right? He's he's um he's trying to warp reality and the more interesting he stuff he has to say and he aligns that to the actual work, the easier it's going to be to get distribution. So you're Elon Musk, you're talking about putting GPUs in space. That's interesting. That's going to find its way all over social media. It that's a very hard thing to do. And again, it doesn't apply to every single business, but finding a difficult problem that is interesting is probably the best thing you can do for distribution. I mean, just think about if you're like an early founder out there, right? You're trying like I'm trying to get on this conference panel and like, well, what do you want to talk

about counterparty risk in CFI? It's like >> you know >> no I mean it's such an interesting point and I think it's actually probably a good segue to like the infrastructure side that I was you know I've like have a couple talking points to talk about because it's like um I would argue that that has uh failed in crypto um over the last two years. Has it failed or have they done too good of a job of it? >> Because you know, >> no. If if you if if if you cannot if you're back, >> right?

>> But no, if it turns out that there's not market pull for the solution that you've then, you know, and and this open design space that you've then created, right? Which to be fair is different than putting GPUs in space. That's like that's a hard problem that you know people already have demand for GPUs. They want more of them, right? And there's benefits to putting it in space. So I would argue that that is the sweet spot of like a hard problem that's really you know easy to talk about and like you can get a lot of attention with something that will have demand right when it's available. Um whereas I would say like Ian layer struggled in terms of actually like the level of demand once

it was ready was not you know nearly as close to what they thought it would be right. Um is this starting to make sense a little bit? >> It does. It does. I you know it's just interesting though if you think back to many of those um yeah I may be a good example kind of talking about refining trust in the same way that was a powerful metaphor that Shriram gave >> oh and crushed it yeah murdered it >> got people so excited yeah >> I mean and think about I think this is why crypto had such amazing distribution in the beginning because it dreamed so big the ideas were so intoxicating and it was interesting to talk about I mean it's interesting to talk about and

I guess it it weirdly the space I think did quite a good job of picking a super interesting problem and then it it got a lot of distribution that way. pill. This is what we called getting nerds sniped in crypto. That was distribution in a way. Y um >> especially for the investors, right? And less so like the demand side of people that are like dying to try these products like and and pay for them right away. Um yeah so like the framework from the infrastructure side that I think about a lot is uh you know there's a good book basically introduces market pull or technology push as the two basic ways that like new business ventures come into the world right so market pull

is like you've you keep you have identified a problem right um that needs a solution in the market or a need for this solution you keep the problem constant and then you figure out the shape of the product that addresses that problem, right? Technology push is, hey, we have this new technology capability, right? Non-use case specific. Um, this is like what layer ones are basically like you've got a feature set that you hold constant and then you go find a market, right, that can leverage that this feature set to manufacture market pull, right? It's like, well, we have this thing called the EVM and smart

contracts. Oh, this introduces a new class of applications. Let's likely start with financial applications, right? And so, Ethereum um was actually quite successful in taking this like technology push approach. Um not with every use case, but at least with early, you know, financial use cases, right? Um, and so I think crypto historically like how as when you're building infrastructure like there hasn't been a ton of stuff that is like very clearly working and there's demand for it, right? Um, until like this past two years with stable coins, prediction markets, RWAs, things like that, right?

Um, so infrastructure has historically always been technology push, right? And the winners um were really good about identifying like those markets and being very hands-on about getting builders to manifest like what's possible in those markets and more recently it's kind of fallen flat right like I would say a good example again not to pick on I layer I love them I'm an adviser to layer but they had this as framework right which opens up that's technology push now we can have you know trust services for anything offchain with no like limitations, right? And turns out that there is not as much market pull

for building with that infrastructure as they thought there was, right? So now they go find what they believe is like where market pull will emerge, >> you know, most uniquely for their solution, which is the AI side, right? And they try to pull that vision in and manifest market pull that way, right? Um, so I'd say like from my st in my seat like leading ecosystem development and growth for a new layer 1 that is super novel and unique and there's so many things that you could build on delta right um just putting out a general purpose framework and saying like here's the design space unlocked come build anything like that's not

going to work anymore right um and so the ones that are I think finding success is like okay what do we think is, you know, uniquely enabled by this by our infrastructure. Stack rank those. Let's go not only like put out a general purpose SDK, but actually build tooling around it for those for the ones that we're taking a bet on, right? And then we'll iterate um and go from there. And I think like the tension is that last point. um you historically have wanted to present your infrastructure as like the biggest possible thing to you know investors or users right we call it like biggest folder you know mental folder

theory like you want to be basically thought of as like you know the first thing since Ethereum rather than you know Ethereum roll up ZK rollup then you're like that's where your folder is right and there's definitely this is at odds with picking an opinionated use case where there's market pull for today, right? You go like I think you still prioritize market pull, but you have to show investors how this wedge expands out into something that is massive longer term. Um, so yeah, that's I think it's a very interesting point to be launching crypto infrastructure because the stuff where you just put out a grand vision and this thing's going to

be huge but you can do anything with it, that's not working anymore. Um, so you have to be narrow, but then that doesn't that makes it harder for people to see a token price that should be valued at 10 billion, 20 billion or whatever, right? Um, so yeah, anyways, you've just got into my brain as [laughter] an L1 devel, you know, operator now. You know, you know what else is a now I don't think people necessarily thought about this, but to pull back on to call back on something that Zave pointed out, one form of distribution as well has been the the reflexivity of token price is a huge form of distribution. Um, and so I think Bitcoin

and Ethereum kind of organically stumbled into that, like the the price going up so much can be a huge attractor for you. And then I think others got more cynical around it and sort of figure that out and then tried to manipulate the price of the token to generate demand, which is tough, right? And so, but if that but if that has been your distribution strategy, if that has been a part of your distribution strategy or if you or if you know, let's not say it in in let's just say you're w you're a marketer at one of these big protocols and you know your token price is going down and you see that onetoone correlate with your activity on your platform.

>> What do you do? Right? >> I get it, but it's [ __ ] broken. Like that is broken, right? like the revenue meta killed the, you know, token narrative strategy basically, right? Um, and the harder you try to lean into token go up mechanics, the more the market hates you and the worse the token will do. But I don't know, maybe I'm wrong, maybe I'm being too cynical. >> I have a comment here though on it doesn't work anymore because I think it worked with Hyperlid. Hyperlquid made the token go up and they also built a product that makes a lot of revenue and the token now is doing really well. It has multiples similar to what we see in fintech. So I think you can still make token go up and you can still distribute

like that. You just need to have very clear PMF to be able to do that. You can't you can't be too early with token go up and having no PMF whatsoever. That's been destroyed. So that's why it's getting a lot harder for infra projects right now. I think Mars you made some really good points before around the push versus pull. Maybe just to like expand on that as well. So I think a big problem with infra sometimes on the application side I think people think very hard around the market and you know why they'd come in in the first place how you access that market. On the infra side you know I've also been in this bucket before in the past as well.

It's like quite easy to be nerd sniped and thinking about the future and how it all changes and how everything's unlocked because that was the original Bitcoin and Ethereum vision and that did happen like we did enable some really cool programmable use cases. I would argue with Ethereum maybe though like they did solve a problem at the time and which we've talked about in previous episodes which is like Bitcoin was programmable or Bitcoin was money but it wasn't programmable and I think there was a market for people already using fintech apps or banks they weren't happy with and developers building applications that were financial that just didn't serve their need well enough and so like there was already latent demand for something like Ethereum and that they were the first to come out and solve that and they built distribution through the token like we just talked

about. Uh I think in this day and age with infrastructure it's really important to think about I do think fortunately or unfortunately depending on who you are in crypto I think we're actually about to enter a new infra stage in in crypto which like it's great because now we have apps um which obviously I'm very bullish on on the infraite I do think we have some emerging problems coming about because of agents there's more volume more traffic more spam like this is hard to handle and like that creates opportunity to build infrastructure around and also with AI now we're using crypto in a new way. So we need to have rails around how we do that in the first place. So there is for me like obvious opportunities.

You can still build infra. I've got no idea how you get distribution around that because I've thought less about it. I think a lot of distribution in infra and crypto is still B2B. It's like go out to the team the best applications axiom poly market they probably share a similar problem. Build infra for that problem and they're probably going to flock to you and then maybe you can capture that with your own network later on. I think that's still a good play if you can do it like that. It's getting harder to build info like that, but I could see that as still being the winning strategy on the distribution side, whereas like on the application side, I think you need to be a bit more creative around it's not B2B. It's like where the hell are these people in the first place and how do you get them

paying for your product? >> So, couple quick things. One, I just wanted to point out this is a so on our a dashboard, the blocks data one, they we just chart the price of ETH and a one over the other. Um I would say like I agree there's like a fine line here right in between uh the the forwardleading indicator of token prices the impact on business activity is so undeniable. I I I kind of understand why people focus on it to the degree that they do. However, if you have if there's very little way to understand this at the time. If you're doing cheap tricks that are sugar high ways of making your token price go up and it's going to eventually go down, then that was bad.

If you are doing legitimate investor relation storytelling, bringing investors on board and then delivering, then it was good. It's very difficult to know in the moment, but I will say it's it's >> it's our favorite quote, uh, violent delights have violent ends. Um, right. Like I I think exquisite taste smile at West [laughter] >> Westorld. Uh, no, I think that's right though. It's like um, okay, you have real PMF, then you have, um, a storytelling. You have evidence for storytelling, then token price goes up, and then I think that starts the reflective cycle. But you can't do anything without people using your product.

>> I Yeah, I I I totally agree. Wait on the I wanted to also comment on Z's point on um infrastructure challenge and actually outline another challenge and maybe highlight why it's been very B2B in crypto and then I'd like to get your guys take on how you think um distribution changes in the era of AI because I think it changes massively but one of the challenges in crypto that exists is if you want to acquire users it's not even the best way to do it right in in a web two You can advertise on Facebook and Google. There are these massive platforms where I can put $10 in and get a user out for the most part.

Now, that shouldn't be your only growth strategy. If that is, you're screwed. But that doesn't exist in crypto at all. That has not been built even a little bit. It's actually a massive challenge on the distribute. It's it's why I think to go back to Z, those examples that you mentioned around it's all piggybacking on other people's network because there is no crypto network. I think that's also why people have tried for so cryptoocial so much. Um, and it just has because everyone's like, "Hey, I got to find a way of acquiring customers." If there isn't a social network or a way of finding from like a Google or Facebook equivalent, it's just very difficult to figure out how to do that. Um, and especially if you're a marketer, you're you're you're at a massive disadvantage.

Um but yeah, I wanted to point that out because that's just a huge difference between user acquisition in crypto versus not um in basically any other business. >> Yeah. And it also like touches on something that we haven't explicitly like mentioned, but there's like a lack of repeatable playbooks that you can run in crypto. like we've said multiple times today like you know I don't think this would work again but it worked for them and like at the time um and yeah that also speaks to like you know like what is a repeatable like successful playbook outside of crypto like very like ad networks like they work pretty well for the most part um but maybe we just haven't tried hard enough like when

I go through the Denver airport and I see like um you know uh an alt layer 1 poster I'm like, who is this for? [laughter] This this is probably the wrong idea of like me, you know, meeting people where they're at. But, um, I don't know. LinkedIn seems like it's actually a good channel or at least the stable coin stuff, like more fintechy stuff. Uh I think the only repeatable playbook in crypto which is starting to break is founder big social media personality talking about ideas getting on the podcast and conference circuit. That's dist that is the playbook for crypto distribution. I mean there's a reason Blockworks has

podcast and conferences. It's like that's how that's what we recognize as being the rails of like you want to build a reputation and early product market. Like that's kind of it. Um that's what you >> And is that even true for like um I don't know like the RWA DeFi stuff where like I think we were before the show we were talking about how do you get like close to capital right um and get those like you know onetoone relationships is going to take a long time but like Apollo will you know deposit $500 million into into Morpho or something like that and it's going to stay there.

I think that's a def let's let's let's unpack that assumption in one second but I think that's controlling a scarce choke point as a distri I think if you are in RWA land right now that is your distribution strategy you it probably matters far less than it ever has that you're big on CT or whatever it probably matters far far far more for you to gain proprietary access to the couple of big asset managers that are looking at crypto because everyone in their mother wants especally to yield at the moment. So if you have some kind of proprietary access or relationship, I I think that's that will be really enduring for at least a couple years. Um but we should talk about AI and can I

can I introduce the mental exercise that I have been so we talked about these different frameworks for AI right which is controlling this choke point. um is piggybacking on existing networks. Um it is building verality uh into the product itself. I'm c the way that I've been trying to approach this problem or this idea is is is kind of a thought experiment of imagine if you could collapse the friction of creating any product or service down to zero. What do moes look like in that world? Like that's my I don't know if AI is ever going to get that good, but in theory it could,

right? So the friction of creating a product goes to zero. How do you create a successful what does that world look like? And then how do you succeed from a distribution standpoint in that world? How do you get a successful product out there? I have thoughts, but I'm curious what you guys what you guys think. >> Yeah. Yeah, go for it, Dave. >> Yeah. No, I think it's a very timely question and we've spoken a bit about the crypto and AI uh craze that's going on now and I think it's always worth touching on every episode because so much is happening per week. I think I've never spent so much time behind my screen as I have [laughter] the last week with just everything going on and how crazy it's all getting. So, I think now if you're a builder, you have to be

thinking about agent native distribution strategies. Um, I think in the past what's worked, again going back to the creativity around distributing the product, you now need to be creative around agents reaching your product. You're not really reaching humans anymore and getting in front of them. You're getting in front of the agents. So, you need to be thinking about where are the agents hanging out and how do you get to these agents and you know thinking from that point onwards. Well, in general, [clears throat] there's kind of three things actually where I'm seeing it right now and I'm still developing this kind of framework, but there's three areas I think startups can play in. And you want to be discovered by agents. So the question is how do you get discovered? And it's not like humans where humans make decisions in different

ways. I think agents are much more rational. So they just want the cheapest, the fastest. We saw that web 4.0 thing get released yesterday where the agent has to survive by continuing to remain autonomous and pay for its own services and then they can replicate etc. I do think we're going into a world where agents will have their own minds and replicate and build businesses and earn etc etc. So the question is if you're a human building a business or maybe your agents would you control? How do you get in front of those agents making all the money uh in future? And so I think there's three areas. So one is data. I do think data I've tweeted about a bit this week. I think data is underexplored especially with crypto rails. And I think APIs is too expensive. They're too hard to access.

Agents won't know right now which ones to use. They just want the cheapest, fastest, best, and probably most relevant. So it has to be contextual as well. And they need to serve at the right time. So I think data is an area in crypto which I could see absolutely exploding in the next 12 months. Another area which I think is relevant to the sort of agent native protocol thesis or framework I'm coming up with is on the execution side. So we work with a team Oya. I'm also speaking with some other teams on like how do you get the agent to execute something on chain in the way that you expect and that's actually a really really hard problem and it needs to be solved in crypto on the infra side. And so then the question is then what why would an agent choose this info

in the first place for how to execute something? Is it the policies? Is it the compliance? Is it how much money they make? Like we are assuming right now what agents actually want. I don't think anyone has a clear thesis on like why agents would choose something in particular. And the third layer is intelligence around like when agents need to like evaluate other agents or evaluate the risk of doing something, how do they do that? How do they get smarter to take that action at that time to achieve that goal? So I think poly market pal all those things actually fall into these categories and I haven't gone deep enough to actually explain in which layer each of them are but they I think most apps will eventually turn into almost like you know MCPs which is how models access data at the time they

need it and then that therefore that is where the sort of value occurs and that's where the business will go and so I think now applications need to and infrastructure to some extent they need to think much more about like building docs for machines and not for humans building their SDK to target agents and also thinking about new distribution channels like you know if I'm doing something in claude code today like how can your project be the thing that I execute my intent with within something like claw code where everyone already is everyone's already comfortable using it or something like open claw where like all of a sudden you have a skill this goes viral and you acrew a lot of value like that I think it's like very very new and people like I haven't seen anything yet which is like well I have

seen a few things which maybe we can touch on [laughter] another time but Like I spend a lot of time thinking about this like these kinds of distribution strategies a lot of them are just like agents serving people's needs within LLMs and that for me is where the crypto space is going in future and so I think it's really important to think now about like how to attack that and just to finish so another thing I think a lot about right now is like we're speaking about distribution and we're speaking about AI well like chat GBT and you know Claude code and whatever Google Gemini they already have a lot of distribution so The question is like why can't they eat your lunch? And that is actually really really hard question for a lot of entrepreneurs to to answer because a lot of the time when I speak with

entrepreneurs they're trying to pitch me something which is kind of like rails for agents stable coins etc. We've spoken about it in this podcast already but realistically you know stripe chatbt all of them are trying to push in that direction. They're trying to own all those economics and capture that value. So you need to think of really novel ways where these incumbents won't play if you're going to build in this agentic native distribution space. So I I think of things like compliance maybe like domain specific agents which execute things crypto natively that you know um incumbents won't touch. Maybe connecting multiple LLMs and being more open a bit more like open core and multiple agents across different domains. I think incumbents won't do that. And also just the trust layer in general around like how agents actually access and use and

how we trust these things to work on behalf of us to achieve the outcomes we want. I think these are all places incumbents won't touch. So if you are a startup building uh towards agent native distribution and you fit in one of those buckets across data execution intelligence or across those things I just mentioned uh I do think there's a good chance that you could be a big winner in the space in future and then for existing ones they need to think harder about how they access these agents. Yeah, Dave, I think you covered it really, really well. Um, and like we just talked about what has worked from the distribution playbooks and what hasn't in the past. It's like questionable how relevant [laughter] those like lessons are. Um because I think it's pretty fair to bet on like whether it's B2B or even like in BTOC

settings like you have to shift from like human centric optimization for sales to agent ccentric and like there's some very basic tablestake stuff which I don't think will be that interesting in six months which is like making all of your docs readable to agents making all of the APIs open and with the X42 compatible like all that will come quickly. what are like the hacks, right? Um, and like Moltbrook was interesting, right? It was like the Reddit for agents. I was like, well, kind of just want to deploy my agent to go shill my product like to other agents here, see what sticks like it, you know, there's no downside in that. Um and yeah, I think that's the challenge is like uh

that you touched on as well is just like you can either get Open AI or Claude to you know like fold your service into you know their their product um in a very coordinated way. But if you're outside of that like if you don't have that leverage to to get their attention um you know you got to be more creative right and you got to be find like more defensible modes um cuz yeah I think like the dream and this is already you already see this like with commoditiz services is like do you want to use superbase for your datab like that purchase was made with a human approving the agent's recommendation >> um and if you're in that position you're

going to be a rocket ship. Um, but yeah, it's unclear how like easy that is to do outside of like commodity more or less B2B tech products. >> I'm curious. I mean, as a, you know, crypto data company like Zave, why so for instance, if you could look at all of these LLMs as they're really aggregators that sit on top, right? If you're the input into an aggregator, that's almost never been a good decision. How would you think through the strategic rationale of leaning into that?

>> I think it's a brilliant dollar question. Um, so I think there's two possible futures that that might happen here and I'm I'm 50/50 on which one actually happens. So there's one feature of agents where like agents work for humans and you know we it grows through humans kind of having the not the final say but the control over the agents and helping them achieve the outcomes they want to achieve and serving the human not serving themselves. I think the second future which I'm also 50% chance of happening is that the agents serve themselves and they don't serve the human and they do whatever they want to survive basically and make money and continue replicating like life. And in that scenario, you know, all of our assumptions about what

a human wants with data and how a human makes money like an agent might have a completely it's a different market. They might think like the data is either not good enough or it's too expensive or they just think incredibly rationally. I think they will and I think they'll be profit maximizing. And so then the question there is like what do they want to maximize profit on and how will they make money in the first place and these are like very very early questions I think right now which I don't think any company's fully worked out but I do think that like for example blockworks data like how I would use blockworks data is like to inform my own insights about the market my own research maybe to invest it's very like rational for my own needs my own purpose for an agent it'd be interesting to think about like

ignore Z and what Z's needs are like I'm an agent and I want to make money like what does Blockworks give me that nothing else can give. And that's this is where I think actually like kind of private data becomes like very valuable all of a sudden again and I've been thinking a lot more about this and we have a few portfolio companies actually building this direction around like I'm now facing a lot of issues with like paying for APIs and all these different keys I'm holding and like how expensive it is, how annoying it is to coordinate and like my applications are sort of being bottlenecked. I think solo entrepreneurs can build a lot of cool things right now, but with a lot less money, but you still need to pay for certain things and like I don't think the business model around how you pay

for APIs with agents right now is solved yet. That in itself is a crypto opportunity for me. If you can actually build around that and then maybe that's why agents would hit you a bit harder in future. But yeah, I I do think this is all changing in in real time and you know, I'm curious to hear from you, Mark, I guess, on the blockwork side. Like do you think the agents would be rational? Do you do you think they'll make money? Do you think it'll be humans controlling the agents? like are you building pool of this future already or how do you see the block works >> or or do you build the agents yourselves right or like do you create those interfaces yourselves that that that avoid you know getting aggregated right >> I think that that's the the billion dollar question actually for me is who

owns the interface and I what one thing that I frankly is just a an unanswered question is are you going to go to chatbt or claude like a god and and everything like you used to have brokerage and you used to have social media and but it now this thing is just so good at everything that you just go to that I I I um I would bet against I don't know actually I I go to I already use these services for a lot I I don't know uh but a huge part depends on that right if if they're truly aggregators across everything that's pretty tough uh that's pretty tough to see how anyone really does that

well in that world um because they'll have just an immense immense immense amount of leverage. Um but I I think there's a very plausible future where um where the human mind still just needs to separate things out and I go to this for this and I go to this for this and I go to this for this. Uh in which case I think that you can basically okay to to rewind to my original premise. What happens in a world where there's zero friction for creating new products?

There's a gajillion more products and probably the difference of quality in the product falls. So where I kind of go back is I think distribution is more important than it has ever been in this world. Um like I think that in the world that we're headed to, if you play it out not even that many steps ahead, if you are building with distribution in mind, you're cooked right out of the gate. Um and any sort of attitude like build it and they will come is just That's never worked, but it's doubly not. It's 100x not going to work here. So, what do you have? You I think that finding a way to durably get agents to favor your products. I agree. I think that has to be one lever. I think speed

is a big lever here. I think the faster that you can figure out actually one area that I go to is think that, you know, I posed this question on Twitter the other day. Why isn't there 100x if agent if LLMs are as good as everyone says they are including me and I use it all the time. Why don't we have like 100x better content and more content all over the place? Where is it? Like these things are god. Like you have 160 IQ person that will work 24/7 around the clock. Where is all the good content? I I don't know. But it hasn't been cracked or figured out yet. But whoever figures that out is going to do disproportionately very well very quickly. It's like these are weird examples to use. Hitler figured out

radio. Like that was an enduring ve like important advantage. Um you you see all these examples of like early YouTube like Mr. Beast really cracked the YouTube algorithm code. Trump figured out Twitter. Um arguably Twittered his way into the White House. I mean leveraging the new medium that exists for capturing attention is powerful. It makes sense to me that LLMs would be that next thing, but I don't think you have forever. I think you have a little while to figure out how to do it. >> Um, and then >> do like Sorry, go for it. Go for it.

>> Sorry. One last thing in in a world of product is I would products being infinite laws are still a thing. Um, I I would actually predict that lobbying, but maybe the hyperlquid news, this is why this resonated with me. Doesn't matter if all the law, you know, all the products are equal. It's illegal to use that product. you have to use this product or this product created in this certain way because it's safe. I I would I would anticipate a 100x increase in lobbying in this world. Um, hey, we can't compete here because everyone's got the same product, but going to go make have the government say you have to use this one. I I I know everyone hates that as a moat, but I think it's pretty rational the way that a lot of these companies are spending.

>> I think it'll be the same as crypto though, won't it? Where like it doesn't matter if you try and lobby agents will be open. like open claw is really proving that people like open agents working across their tools and even if you try and block it they're going to find a way to circumvent that I think so you can do all the lobbying you want it's like finance and banking doing lobbying against crypto and then eventually gets regulated and people can use crypto including those institutions so my guess would be that AI will go through a similar phase of people will lobby uh you know people will find ways of getting around the lobbying maybe through decentralized and open rails and then eventually it'll be regulated and you know people will know how to use this stuff. So I do think some a prediction I want to make here which

could either be very right or very wrong. I think open claw is showing us right now how people will use AI in future which is really the agent knows everything about their needs what they want to do how they want to achieve greatness or productivity or whatever it is and I think let's say I'm a data platform right now it's still quite human in the sense that these agents are kind of being fine-tuned by humans but the way that humans are accessing these platforms are completely different. So the way that like I'm hitting these APIs right now is complet like I hit X APIs for an application or my open core agent does for example but like the way I hit the X APIs is completely different to the way that I use X on my feed for example. So it's a different way of

adopting and accessing the data in the first place. And so thinking from first principles around like okay human does this agent does this this is what the human wants from that agent in the first place and like how can I then like move my product towards that direction. So there's a lot of outstanding questions. Maybe we can speak more about this another time, but it's also exciting and I couldn't agree more with you, Mike. Uh distribution has never been more important in the era of AI. This one's uh specific around media, but I have a prediction here, which is let's say let's say over time LLMs and AI get so good that you know right now we're kind of riffing and we have these opinions, but let's say in the future we're all going to have our Mike's personal agent and Zaves and Mileses and

you're going to ask before the podcast, what does the agent think about this? And he's gonna give you, [laughter] you know, basically the same stuff. like I I could see the the quality of content converging because everyone's referencing the same source material. Again, how do you curate in that world? I think it has to do with brand. I mean, I I actually think you you probably have a uh there's a brand halo effect that again, if the if the quality all converges, it's kind of like with no brand, AI equals slop with brand, it's like 100x leverage.

I think you've got maybe six months to really build a brand. Um, and then it'll be much harder to do. Um, >> interesting. >> Yeah, I don't know. I feel like the best content is stuff that's like written about like on a hunch, you know? Like I have a hunch. It's not there's not enough data and evidence to for like an AI to come up with this to like put together a coherent story, but it there's like a h like a human hunch that has like legs to it, right? That will never that's not going away anytime soon or getting commoditized. Um, >> I have I I have a slightly different take actually on that which is >> because that's just >> I think they'll be as good as I think good content weirdly comes from imperfection more so than >> Exactly. Exactly. You will be wrong but

people will agree with like the the essence and the sentiment the attitude that's like you know leading to that hunch >> maybe. >> We'll see dude. I don't know. I mean look I think a best case scenario for the whole world is that AI has a last minute last mile problem on almost everything. That would be the best case scenario, right? I mean, you know, it because it is it's hard to create a super compelling bright future if AI just solves everything perfectly. That's >> a tough, you know, great if everything got to the last mile and then it couldn't quite figure it out. That would be awesome.

>> I I think one point to make which is relevant to what you guys are saying, which I feel quite strongly about. I think humans still want to see and read human things and so it doesn't matter how good AI gets and there is going to be some great AI podcasters and videographers and stuff like that but like when I read things that AI generates and shares with me it's it's too perfect. It's like it's clearly not written by a human in my human voice I'm not like this is like a great like a person talking to me like I'm not resonating with the writing as much. If you post an article on X right now that's just completely generated by AI, people are going to work it out because they would never say in conversation the way it's being expressed online and then you look like someone using AI and

therefore probably less intelligent in the first place. And people have a natural instinct for this. And so I think even if AI continues getting better, like a lot of the best brands I think actually going to feel the most human, which is actually interesting to think about. And how do you like leverage AI in a way to continue, you know, being creative around how you build that brand, but then also not so much where it's just like copying and pasting everything from AI, these videos, and I mean, some of that will work as well, but I think it'll be a 50-50 split again here. But yeah, who knows? I think it's early days. >> Do you think that that's just a parameter change that an LLM will eventually get better at? It's like, hey, this sounds too much like it came out of an LLM

>> or or even make it right like me, right? like, "Hey, by the way, I've written a ton of stuff. Here's here's my style. Write it in my style." >> I've tried doing this multiple times and I give as much context as possible. Like, here's my writing style. Like, generate this and I I am actually unsatisfied. Coding on the other hand, I'm like prompting and I'm like, "Holy, like, I've just built like 10 apps in one day. This is sick." But like, when I'm like, "Here's me." And go and write something in my voice or tone, it never comes back with a good result. Because I don't think AI has that human instinct yet. when AI has a human instinct, maybe they can start writing in a way that's more interesting. But at least now it's too early.

>> We're diverging here, though. But is it possible that it's that's still the same last mile problem, but you are far more attuned to differences in your writing style versus you're not an engineer by trading. So maybe an engineer might look at the apps that you've coded and they're like, "Dude, this works fine, but it's kind of dog." You know, like >> it's the same It's the same idea though. Like the bottleneck is the review, right? Like >> um Yeah, >> the it's going through the code and being like, "Yep, that all looks good." Or like surface to me like anything sketchy, me going through some, you know, blog post and being like, "Eh, get rid of that m dash." Like sound more conversational here, right? I think I think I think um and there's going to be

like figuring out where that last mile will always persist versus where like full automation um you know can run free is like a big question that you know we should all be thinking about. C can I give you guys a I was talking to someone at Blockworks the other day and he was mentioning that I think everyone has a inclusive of right management right uh that this idea that AI can't really come for my thing and then you see it come for your and you're like oh boy it actually can do that pretty well um [laughter] but I I think that there's also another element to this maybe this will resonate with some people listening uh if you made the transition from individual contributor to manager because as as an individual contributor,

you're doing the work and you love doing the work, right? Like you're let's use the example of a journalist versus an editor, right? Journalists, you're writing, you know, you got the creativity, maybe not like a lot of journal, but maybe like an op-ed writer, right? You're you take pride in the work and the work is what gives you fulfillment. As soon as you move to a managerial like a manager position, you're not actually doing the work. What are you doing? You're take that m dash out. Um, this sentence was too long. Uh and then by the way you have to you know >> you're setting strategy you're setting strategy and then reviewing right and and that's what everybody's like role is shifting to right now >> and most people when they make that transition don't like that it's less attract it's less fun right I think the

good managers do it because they recognize that by cultivating the skill set they can have a larger impact but it takes a long time for that transition to so I I I have this like little bitty theory that every many people are understand they're like wait I don't write this anymore I don't think of these podcasts anymore I don't do the podcasts anymore I just prompt the thing and it does it this doesn't feel as good uh and I so I think there's actually a equal parts like I think a lot of people are worried about job displacement but it's also like [laughter] it's like making the transition from IC to manager in terms of the joy that you take in your work which um >> exactly well let's see if people notice when use like Sora for the next podcast.

Uh or like some week where none of us are available and we'll just see if they notice. You know, we'll we'll give it the prompts. We'll tell them to run wild with, you know, we'll give them some head shot and um let's just compare and contrast. >> Wait, can I give one more example for why I kind of don't think that it's all going to be on like chat GBT is the god mode box. If you look at Claude, so Claude just released. So actually using the example of Sora, remember there was at chat GBT there was that code red notice a little while ago and they moved away from Sora and product just to research. As good as these LLM and these companies are, you can't there is some law of attention at a company and you can't do everything well. Problems are

hard to solve. You're bottlenecked by the human capacity for attention at a business. And I I think that that law still applies. The other thing though, if you look at Claude, um, a lot of their integrations allow you to bring Claude into where you already do work. So, Excel, I don't know if you've been watching, they keep they keep making it easy for if you want to do calculations on Excel, you can pull this in and this in and this in and this in. So, that to me signals they're building for a world of multiple interfaces. But as soon as they are the ones that get aggregated right by by openclaw or whatever it's called now they they send you know they shut off the APIs they send a cease and desist and then that person goes and

joins open AI right and so I think like that that is also another example of like if you know open beats closed um and that you know gives me some hope >> I think I think open does beat closed and I think we're seeing this in real time open closed was a clear example of that bought by Open AI. There's going to be several more examples and at some stage when they can't buy anymore, they get sick of open winning. They're going to try and shut these things off and then that's when open source AI is going to really shine. So, yeah, interesting times.

>> Awesome. All right, guys. Oh, were you going to say something else? >> No, just that we're an AI podcast now. >> We're an AI podcast. >> Used to um No, that was a lot of fun. >> Uh that was good, guys. All right. Well, hopefully there was some tactical takeaways there um for gaining distribution and uh yeah, happy happy miles. Enjoy the rest of your time in sunny Denver, I guess. Uh >> I will I will um and yeah, we'll do a live one what in like a month at for for D just to give >> Yeah, we've got D coming up. Yeah, thank you. Daryl did my duties here for uh for showing.

>> Yeah. Um we are we're on we're on track for it to be the biggest one we've ever done. Um and yeah, we have a bunch of great institutional speakers, heads of SEC, CFTC, um all sorts of stuff. So yeah, if you are building and want your own distribution, plug in to DAS. [laughter] So lame. Sorry. Um uh All right, guys. This was a fun one. Uh, and yeah, we'll see you guys both in uh IRL in uh >> let's do it. Cheers, guys.

>> Bye, guys. >> See you guys.