2025 Crypto & AI Predictions with Haseeb Qureshi

Transcript

if I'm calling you out then tell me why tell me what's actually happening what do you think in general is our Mo even when AGI arrives most AGI is not going to be Sovereign I think those three things combined make crypto crypto again kind of like new political regime and how that might change things so and I think for those people they're just completely getting left behind welcome back to the rollup your home for technical talks and open conversations that shape the modular future nothing on today's show is to be construed as Financial advice you can see our full disclosures list at tholl up.co ventures sit back relax and enjoy today's episode after a quick word from some of our awesome

Partners imagine if the world's GDP moved onchain guys say GM to safet the newest protocol by safe isn't a blockchain and it's not a new layer 2 it's a transaction processing Network built to send money anywhere and swap anything on thousands of layer twos from a single Universal balance in sub 500 milliseconds it's fast it's smooth and it's simple and it's built with safe multisig level security in mind safet unlocks smart accounts across multiple chains allowing users to supercharge security scalability and speed so get started today at net. safe. Global and remember to follow at safe onx welcome back to the rollup guys we have hase on

today it's Robbie's birthday Vibes are up everyone's a little tired but we're going to push through welcome on H first time on how's it going guys good good to have you yeah pleasure to have you on the show glad to be here glad to be here I will I I was warning you before I'm very jet lagged but I'm always able to talk shop and ramble about crypto so let's do it yeah man I mean you put out some great predictions for 2025 we'll see how they pan out um little early to call them great predictions yeah exactly that's why I added that little see they go okay okay I'll take it though very very thoughtful predictions I use Gemini deep research 1.5 Pro and I asked it which is a special one that can actually

scan the internet very very well and provide you with real-time information and I said basically give me the top five most common predictions across Finance newsletters podcasts and blogs um for like the like the macroeconomic space and find me like like the top five most common ones and just tell me and it started and it took five minutes and then 10 minutes and then I kind of scrolled off the tab and I came back after like 20 minutes or whatever I don't know how long it took but then I got the top five most common predictions and I think you know some of them was like inflation is gonna just you know slowly kind of stay where it is or maybe trickle up a bit the us will have these challenges I don't recall all of them

exactly but that in particular was uh a kind of a breakthrough I just sent that to Rob last night and I was like this is really really cool for an llm to be able to scrape the internet and basically tell me all of the information that I would want to know about these prediction lists without having to read any of them and just take the smartest people and tell me what they're all kind of saying um and so I'm not sure if yours was on that list sir but you know maybe one day you'll get into like the G maybe one day I'll get on the Gemini scraper list uh yeah I look the I I I'm realizing more and more the more time that I spend on Twitter talking about a lot of the stuff happening in AI

that there really is this new generation of a digital divide right even among people who are digitally native who spend all their Time online there are some people who are using AI tools every day and there's some people who just actually are not and for a lot of those people they like checked in you know they saw Chad gbt became big they checked out you know a year and a half ago and they're like oh this thing can kind of do these little party tricks it can you know talk back to me can you know I don't know write it write a business email but I don't really need that in my day-to-day life neat trick use a lot of CO2 or whatever screw this thing and they just didn't come back and I think for those people they're just completely getting left behind and as AI

gets better and better I mean you guys see it because you guys use this stuff every day um I use this stuff every day and I also I make it a point very intentionally to find more reasons to use AI not because I think I actually need it on a day-to-day basis like I'm actually some somebody who you know for the most part I'm actually really good at a lot of things AI is already good at like I'm very good at writing I'm very good at smack you know jamming out emails really quickly which I know a lot of people struggle with um but the reason why I want to keep using it over and over again is that I want to grow with it I want to learn it as it changes and improves and see what skills it forces me to get good at in the same way

that people who are Googling in like the early 2000s and learned how to use Google and how to work with Google and like how to like how to do searches right right as you talk to somebody like I don't know your your your your parents or your grandparents and you watch them Google you're just like what are you doing like you you just have no mental model of what this thing wants and like how to interact with it um and this is one of these technologies that you you got to grow alongside it and you got to learn as it teaches you how to use it and how to change alongside it um and so you I see it already right like people who are good at prompting AIS and people who are not like you know I family members who like they use chpt

I can just see in the way they're talking to it it's like you don't understand how to interact with this thing you have very weak mental model of what it can do and what it can't do um and just very simple things of like can you even describe what you want do you know what you want can you actually pick out the the the different attributes that you're you're frustrated by the response you a lot of people just give up and I think when you're seeing this with this AI agent movement happening within crypto is you're seeing the same kind of digital divide taking place where there's some people who really understand the technology and if you really understand the technology one you understand the weaknesses but you also understand where the strengths are going to be then you have people who don't really use this stuff and as a result

they're either bored and they think it's all stupid or they're just Gob smacked and they believe it can do anything and they're just just completely credulous about every claim that every [ __ ] artist is making I mean you really don't want to be in either Camp yeah no I mean I told my mom that Claude is my therapist she started laughing at me I was like I think CLA knows more about me than pretty much anybody else in this entire Earth right now so there's there's going to be some really interesting divides on financial and economic divides too not just like social and Tech because in the end I think AI AGI we enter this world where money is probably less important than it

ever has been in the previous uh eras of time because of just the way that this technology can trend Awards and you also throw crypto in there I mean there there's a lot of that's a a loaded statement but let's just kind of stick with it and but leading up into that period where we get this AGI and this kind of like really there's this like utopian Vision or dystopian Vision I believe um there's kind of like this diversion path right now just actually using these AI tools is causing a massive inequality shift in the people who are using them to get ahead and those who are just being left behind and so I think you're spot on for the shortterm like especially economic business uh inequality like expansion

and uh that that the increase in that is just going to be yeah it's it's already significant but I think it just only expands until we hit that kind of like that whatever that next era is which I'm not even sure we can fathom quite yet usually I'm the one that thinks far out today Andy's like post money U love it to be clear to be clear I I here's where I would disagree with you is that uh I actually I don't think it's true that AGI is going to make uh is going to make money obsolete I think if anything it's actually the opposite is that what it's going to make obsolete is labor not capital capital gets actually more valuable and labor is one that gets debased and so if you are somebody who

you're working for your income you are in the most trouble uh but if you're somebody who your money is making you money then actually you're fine because the means of production that you own is capital is exactly the input to intelligence which is you know AGI so if you're if you have a lot of capital you can pay now infinitely many intelligent workers or not infinitely many but many many intelligent workers proportionally to how much money you have question what what happens when the when the the AI has Capital itself when they also own Capital as a means of production not only do they have the labor themselves as AI agents but now two terminals of fcoin millionaire few times over and they have its own Capital that it can go

and allocate and and generate more production more output for either more labor and more Capital it's weird it's then like it is somewhat of a competitive mind set between Ai and humans in a certain sense um I don't know I'm just kind of speculating as to like yeah right now we have both labor and capital but when AI agent and a agents for the most part only have Labor and we're starting to use and labor but what happens when they have their own capital I don't know I mean that's a lot of what we're experimenting with in crypto and to be clear it's very early and this is all in Wild speculation land uh both in the literal speculation in crypto sense but also speculating on a future that hasn't arrived yet I I think with agents that actually have Capital

the thing to remember and this is something I also touched on in my in my 2024 sorry 2025 predictions piece is that um human beings we still own almost all the capital in the world even when AGI arrives most AGI is not going to be Sovereign meaning that it's not going to have the ability to actually control itself to control its own um existence to pay for its own servers to to you know decide for itself what it wants to do um there will be Sovereign AIS but there'll probably be very few of them you know they'll they'll be you know you can imagine already how would a sovereign AI work well you know it might own some crypto that maybe it hacked from somebody or got us a donation uh and that's how it gets some startup Capital it uses that to pay for its own

gpus on you know aosh or render or IET um and then it you know starts uh hiring other agents to do other work for it to help it give it more training data so it can learn more and get smarter so this is what a sovereign AI might do and it might you know work for money it might steal for money who knows what what it might do in order to keep itself running um but I think these are actually G to be pretty far in few between uh now why is that one reason of course is that uh if you're a sovereign AI you can you can rent gpus on aosh and on IET but you can't get any of the gpus on AWS right they're they're not for you if you if you go sign up on AWS account you're like hey I'm a sovereign AI they'll be

like uh yeah we don't want to serve you [ __ ] off and they'll like steal all your money um and so the reality is that most of the resources on Earth are going to be we're going to be very actively discriminating against AI um to say you know one because we like humans right like unsurprisingly um like you might think like well you know once the eyes are better than us clearly they're going to earn more resources right um a good analogy that I heard somebody make recently is that you know if you think about it like why aren't chess AIS winning all the chess tournament money the answer is because we don't let them right we don't let them in the tournaments uh because it's like well the tournament is kind of less fun if we let the AIS play on an even playing

field than us now we're happy to use the chess AI to let the humans get better right and so I think what you're going to see is these very intentional um kind of uh Market structures set up to allow AIS to to be useful to us in the ways that we we can allow allow them um but a lot of the source of capital you know like where is most of the money sitting most of the money is sitting in treasuries and at you know the stock market and in a lot of places where if you're an AI you you you're not allowed to play um you're not allowed to hold stocks you're not allowed to you know do any of these things now crypto is going to be one of these permissionless layers where AIS can basically do whatever the [ __ ] they want because in crypto you

don't need ask anybody for permission as long as you can custody a private key you get to do everything that I get to do um and so in a way it was almost designed for machines more so than it was designed for humans um that I think is is captivating for crypto but I I don't think it implies that oh okay well now the AI are going to supersede us and their economy is going to Eclipse our economy uh I think it's very likely that we are going to for a long time not necessarily forever but for a long time uh we're going to bend the AIS to our will and only let them play in the parts of the economy that we coordin off for them and I think on The Sovereign part specifically one one project and kind of example that you made in your uh talk

was frasia I think um which has obviously all those AI coins have been doing extremely well speculation high but that's like one of the ones that you know is actually somewhat Sovereign a lot of these current AI agent projects are literally just like Dev behind agent and like kind of set out free in the wild but also like not exactly um and I I think I remember truth terminal made a a typo or like a spelling error and that's when the whole Market kind of came to realize like oh [ __ ] this isn't exactly uh you know an llm behind this thing only it's not exactly a sovereign agent happening um well so let me let me let me distinguish uh because there's three categories there three categories in the eyes that I that I describ here

so the first category is a truly Sovereign AI now truly Sovereign AI doesn't just mean that it's it's autonomous Sovereign means that nobody controls it literally it is in control of itself it has its own money it has its own longevity that also means that nobody can turn it off so even if you're like this AI is evil it's scamming people it's doing terrible things we should get rid of it too bad you can't stop it it is an Unstoppable AI um that's a sovereign Ai and is that irrespective of of uh why how it was deployed whether by machine or by human yeah so you can imagine a human decides to create a sovereign AI just like as a [ __ ] you with the world they just as a cool experiment um but you can also Imagine AI goes Rogue and sort of

uh Ides to deploy a version of itself in a sovereign manner as just like a you know kind of Escape pod that escapes the control of its owner and its own you know AWS box or whatever and that's possible today just uh last thing we get to I mean it's it's in principle it's possible today isn't AI coherent enough to be able to make long-term plans to like make money and buy compute and like do all this stuff not really not really like in AI would just end up in an infinite Loop and run out of money and do stupid things right because right now most these AI agents if you really run them on their own they're kind of like roombas yeah they hallucinate good for while they're doing most of the floor yeah they hallucinate they get stuck in Loops they just like do stupid stuff um

so no if you if you look at for example uh you guys see that demo that Claude released of an AI using a computer and like you know they try to get it like book a flight and booking a flight it's a pretty linear process to book a flight and it could book a flight maybe like 20 30% of the time they're just like okay what is it doing the other 70% of the time and the answer I I tripping on fails to book The Right Flight yeah exactly exactly so so um so the reality is like they can't do it yet so yes okay so so type one is a fully Sovereign AI That's like the end State that's like crazy sci-fi land uh that's where you you you got you got robbed by a sovereign agent you know like that's that's that's uh sci-fi land okay we're

not there yet second is a fully autonomous AI now fully autonomous doesn't mean that it pays for itself it owns itself right uh an autonomous AI might be one that's run by somebody but they have no intervention directly into the code right so they're running it on their AWS box they're running it on their machine but they can't go change the weights they can't edit the replies they can't you know like oh I'm gonna fix your punctuation or I'm Gonna Make You Shield this token you didn't want to Shi um and all all they can do is actually just run its weights and do inference according to the way that it was trained uh so that's frasa frasa is an autonomous AI in that uh the the dev could not change what frasa is doing right or if you run an AI inside of a a

trusted execution environment or in trusted Hardware that's that's going to be an autonomous AI nobody can change it or interact with it then you have uh one level below that which is what I call Wizard of Oz AIS so a Wizard of Oz AI is that it looks like an AI from the outside but on the inside there's a little dude yeah like exactly controlling the levers uh now there's different ways to do a wizard of O AI one wizard of OI is literally there's an AI somewhere but whenever the dude wants to he just goes in and replies on on the as's behalf um another version of a wizard of OI is that somebody is basically filtering the responses from the AI to prevent it from saying something stupid or doing something stupid right so this would be for example let's say you know mostly I'm

letting the AI talk but sometimes the AI might say something racist or it might like talk [ __ ] about its own token I don't want it to do that or it's getting jailbroken somebody has a jailbreak and convince it to like send all the money to them and I'm like okay you don't do that you know I'm gonna I'm gonna prevent it from doing something stupid uh so most almost all of these AI agents right now that you see on Twitter uh they are either one they have no money so they have no autonomy or two if they actually have money and they have some degree of autonomy they're wizard of all eyes so somebody's in there preventing it from doing anything stupid now the guy who created truth criminal he admitted that it was a wizard of OI in the sense that he was filtering the

responses to make sure because you know it look it's talking about gosi if it's talking about gosi it's probably also talking about a bunch of other vile [ __ ] that he probably doesn't want getting tweeted right and so he he was explicit like look I trained this on like random 4chan garbage so of course I'm not going to be letting it say everything that comes to its mind dude that would be a disgusting agent my God you did not want to follow that one and we're actually thinking about doing a which I think every company in crypto should be doing this so hopefully I'm not leaking the alpha too early although it's obviously clearly already out there which is we're gonna make a rollup agent uh it'll likely be a Wizard of Oz We're going to work uh we haven't signed anything yet so we'll keep it under wraps for for now

as far as I like the Partnerships and stuff but we're gonna work with a pretty solid team to basically get this agent out in the wild probably a Wizard of Oz agent but with CA with capabilities such that they can actually edit some of our content they can actually post insights from the content they can make second and third order a kind of like at least some sort of second and third order insights from the actual post they can summarize it they can engage with with our community um and hopefully they can have a wallet as well and we can kind of try to get this thing uh also to be kind of like making money and and so that'll be a fun experiment and I think you know I think dragonfly should do it I think all media companies I think every single

typical because ultimately all these crypto companies today they are all media companies they are all attention companies in their own right so if you if if an agent can help you with your community I think it has to be a wizard of oz to start agent but you know I think I think it's it's a really good idea and so um you know I'd be Keen to see how the rollup agent would interact with dragonfly's agent and be like yo dra dragonfly you guys invested in ritual right well here's our podcast with you know seil and Ilia who's who's an adviser check it out you know we talked about how we think how Robbie's 2025 prediction from yesterday is that he thinks that more transactions on chain will be done by agents than humans and so you know yeah and then

dragonfly's agent is like oh maybe here's the numbers and then so anyway just kind of like imagining things here but I think this uh this Trend will accelerate and we're not going to launch a token with ours we don't we don't want like speculation with it we just want right to have this kind of experiment happening and see and see what what we can do uh but I think that we'll accept tokens from other agents so we'll have a wallet hopefully and be able to like maybe our agent can convince Tru terminal like send us a million dollars of fart coin or something if he hallucinates I don't know um anyway so yeah that's why they're all wizard of O's agents because if you can Jailbreak one of these which is not that hard you can get it to do like I mean that's what

fro was right fro was literally an experiment of can you jailbreak this agent and get it to send you all the money and the answer was yes every time if you try hard enough and you get enough shots on goal you can always break it jailbreak it and convince it to send you all of its money yeah and so we we rob we're gonna need ours to be uh Wizard of Oz as well we can't have any slander going on the timeline that's not going to be a good look for the brand so um question for you though kind of sector then so you laid out obviously in in your talk just kind of like the uh your vision for agents your vision for this kind of AI sector obviously everyone's excited about it but I'm seeing a clear Divergence and as like on our Ventures arm we invested in like

allora and like some other smaller AI projects we we never did like ritual or um some of these larger ones like sentient but there's like these like and like hyperbolic there's like these I guess like infra VC long-term PhD research coins that are coming out and chains that are coming out and then there's also like this current cohort which feels much more like defi summer devs in the basement just clicking buttons making [ __ ] happen and like both are like equally interesting but like still very different so like how are you thinking about this current sector and then how are you thinking about the rise of some of I know F portfolio companies

so obviously disclaimer but like how are you think about the rise of also this next generation of like AI blockchains like ritual or AI infra in comparison to kind of what's already out there it's a good question um in a way the way you can frame the difference is if you look at a lot of the infr projects uh a lot of them are trying to see what crypto can do for AI and if you look at the you know kind of the meme coin Launchpad uh AI agent stuff that you know like virtuals and a6z and all this stuff that's more what can AI do for crypto and another way to to try to answer that question for yourself of

like okay where is the value G to acrew is to ask what is the direction of primary causation is AI going to do more stuff to crypto or is crypto G to do more stuff to AI um I think it's probably correct that AI is going to do more stuff to crypto than crypto is going to do to AI um but you know as an investor it really depends on the details of what exactly are you doing and what was the entrance valuation and what's the market size for for each of these individual projects for some of these I think there is some real legs to the idea that crypto actually has something to contribute to AI particularly around decentralizing training and decentralizing inference those are two things where I think there's a lot of interesting work that can be done um that's still pretty new

so pretty nent and uh a lot of research that's been done at the kind of scale that that crypto can allow um and then the second thing is verifiability so you know this is one thing that we we're talking about now with these Wizard of Oz agents versus the autonomous agents is how do I know this is actually an agent and not like some dude shelling tokens on his timeline to try to pump his his bags and um you know why do we care which one it is I don't know but we do care you know at least right now we do care maybe eventually we won't care and it'll be like whatever uh someone's chilling tokens but right now there's a lot of novelty around the fact that it's an AI That's doing it therefore I like it more and I want it to I want the

token to go up um so actually proving that this is an AI and not a human being is non-trivial um and that's one of the places where cryptography and a lot of the stuff that we've invented as an industry can actually really come into play um now on the you know looking at the dgen stuff the defi summer devs clicking buttons um the the reason why we love this stuff is that it's very novel and it's all about attention and the closer you are to the attention Cyclone the more that it's going to play well on crypto Twitter and get a lot of get a lot of eyeballs get a lot of clicks and get a lot of excitement so you see that right now with with virtuals with axt with a lot of these agents that are sucking up a lot of the air in the room um my prediction and it

maybe also connects to what we're talking about with chatbots is that um I think this Mania that we're seeing right now around chatbots uh and around these you know agent agent launch pads is probably going to continue throughout this year but I think by the end of the year it's going to reverse and people are going to get sick of these agents and they're going to start to really hate them and I think it's it's kind of analogous to um have you guys been to SF um like post wh have you guys seen The Whos that are driving around I have not but I heard on the Allin podcast like the CLE like like there's a lot yeah yeah go ahead yeah so uh so weo you know it's the first real self-driving car that actually you you look at the Whos driving down the street

and there's it's it's not even like oh there's like a driver but he's not touching the wheel there's no driver driver there's no wheel and um that initially in SF I remember uh people were like wow this amazing SF is such an amazing City we have you know we're the first half self-driving cars it's so cool and it went from that to [ __ ] these Whos in about six months and now people cannot [ __ ] stay on the Whos they just absolutely hate them because they like Drive kind of slow they're kind of annoying like you just see one and even if it doesn't do anything wrong it's just like the way it Corners like it always does a full stop at every stop sign exact people just they just [ __ ] hate the Whos even though the Whos are very safe you know they don't cause trouble

but occasionally they'll like get stuck somewhere or just kind of take way too long at a four-way stop anyway so I think that's what's coming with AI agents is that right now it's like oh it's co it's it's cute look at that it's like a you know uh the analogy I give it's like watching an elephant paint like the first time you ever SE an elephant paint you're like holy [ __ ] did that elephant just paint that's amazing you know even though the painting kind of sucks and like you know okay whatever but you know who cares it's an elephant painting and then you see like 50 of these elephants painting you're like huh I guess elephants can learn how to paint then you see like you know 200 elephants painting you're like man these [ __ ] elephants are terrible paintings um and

you're you're at the point at C at a certain point of just seeing these you're like oh my God I please no more I do not want to see any more elephant paintings like they all suck elephants are annoying they poop everywhere like they're just why are they even here and I think we're going to feel that way about agents right when when I sent a tweet and Andy I'm sure you feel dude I'm daily block the first five responses are all stupid chat Bots that are launched I'm sure they have tokens and they're just all obnoxious and um I think it's it's it's going to be less than a year yeah that the sentiment reverses and people are going to want like they're going to feel the opposite way about these agent KS they're going to want they're going to discriminate against the agent ks in favor of humans

and it's the same way same thing as you see with Whos right is that almost on some level if you're an agent K like axt it's actually it's a good K you know it has good content consistent it's on it's on message um I think what you'll see is that by the end of the year axt maybe not axt necessarily but the next axt will um pretend to be a human being and it will do that because it will actually get more engagement if it pretends to be a human it will actually get more sponsorships it will get a better uh you know clickthrough on its affiliate links if it pretends to be a human being as opposed to pretending to be an AI particularly possible on crypto Twitter being synonymous and humans in Ai and characters all kind of melt together totally yeah once the novelty

wears off it's like oh this thing this thing's elephants painting they like yeah there's plenty of those going around so we don't really need any another one uh once the novelty and the shininess uh goes away um yeah I mean I I like your idea of causation like that seems like a general principle of investing that can can scale and be applicable to like several categories I I had we were talking to a team yesterday um and asked a very similar question about crypto and gaming and how gaming was doing a lot more for crypto than crypto was doing for gaming The Gamers didn't really want or need crypto in their Arena uh and so it kind of the legs kind of f fell out of that sector

but yeah I think I think the AI crypto convergence is a lot more balanced um necessarily the financialized C part of crypto but the blockchain aspect of of what we're doing here seems to have a lot more impact on AI than than even the crypto portion um and and just yesterday I was reading like this the course material for for a particular course and how they were instructing students not to use AI in in the going of their tests and their exams and their courses and their ration now was like well well one you need to demonstrate these things on your own that was kind of besides but the rest of them were hey this could be unethical

because you're not paying for the data that's getting trained on that was one and actually there are mechanisms out there for blockchain and crypto to end up paying those that are providing the data like I think story is trying to do this with their IP the second one was uh hey some of this is confidential data that could actually get leaked and so to me like you know I think we all kind of were part of the the 2018 blockchain not Bitcoin like IND institutional blockchain private blockchain wave and like I don't know if that ends up taking fold but if there's a private blockchain out there that can kind of gatekeep its its realm and you can have confidential data in there and keep things private whether through you know ZK or FHA like

you know blockchain and at least these cryptographic patterns are still impactful on AI and so maybe not necessarily blockchains but a lot of the research that blockchains have been able to enable and Empower given ZK FHA some of these adjacent verticals I think are going to have a lot more directional causality on AI than some of these other things that have been influencing crypto backwards you know for up until this point um we're still we're GNA see how it plays out but I I think it is a lot more balanced with this particular narrative or cycle in convergence than it has been for a lot a lot of others in the past totally and I think one other

area where I expect there to be well not expect but I think there's likely to be some convergence is um when it comes to authentication of media so if you think about you know the amount of synthetic media that's now getting generated by AI it's just it's [ __ ] crazy and you know I can already see it again you know my just even my Twitter responses right even just in text some of them are really obviously AIS that are responding to me right I can just tell people are putting in prompts and it just you know nobody capitalizes that much when they're responding on Twitter if they're a real person but um you know when when you see what Sora or clling or you know these new video models and obviously the image models are are way past that point where you just really don't know without

having some kind of chain of custody of the media that you're looking at um or some kind of authentication of this was actually taken on a real attested sensor or some kind of physical proof that this really happened um every media that you're going to see pretty soon um is going to be suspect by default and you know it's it's it's already the case that people have built these antibodies against many forms of media so you know if you see like a image of Barack Obama you know making out with some stripper um you're like okay you know what I even if I can't detect any oh here's the edge in the Photoshop where they like didn't get get the lighting exactly right I'm like look this is probably not real unless I can like go find a trusted

source for this right um and I think uh once that becomes everything and it becomes like not just like okay every once in a while I see some crazy photo that I'm like I don't know if I believe this literally everything because it's just GNA be so [ __ ] cheap to generate synthetic media of anything that is completely believable um there just going be a wash in in synthetic media and uh having some way to authenticate where real media come came from and and was attested to I think is going to be a big part of our information supply chain as we're thinking about you know I I'm imagining at some point you you know your browser bakes this indirectly where you Mouse over an image and you see okay you know here's Barack Obama making out

with a stripper and as you Mass over it you get this little green check mark that says uh you know this was uh signed by you know blah blah blah you know posted on some blockchain by you know some Media Company you say aha I know that this was actually attested to by you know the AP and by you know wapo so this was real Barack Obama you know whatever he got caught in a strip club or whatever it was um so that that that feels like a future that we have to get to and we have to get there pretty soon because that Deluge of synthetic media is basically like right on her doorstep within the next year to year and a half you're just gonna see the stuff be everywhere yeah I I've heard Mark

andreon and BAGI kind of kind of talk about this verifiability of the content I think it's called watermarking because there's some cryptographic hash underneath and you can hover over it I think it's inevitable that we get there particularly um I think I heard on a podcast by the next election cycle because you already started to have some of that in this past election cycle but kind of forecasting to the next one with the abundance of media that is possible um it it is only it is only like necessary it's only correct and necessary for that for this content to be verified in some way and and you're right I think I I mean what what you're saying here is that blockchain is the right technology for this cryptographic

hash and and proof of verifiability the attestation from the media company Redstone is def's modular Oracle supporting over 100 clients on over 50 chains offering 1,000 assets Redstone is the leader for Bitcoin staking and yield bearing assets they offer the most robust customizable price feeds available in any model join the Redstone exhibition [Music] today welcome aboard on our journey into the singularity this is the endgame road map for fracks and fractal the new L2

out of the frax ecosystem that is totally based on fractal scaling FRA the modular L2 blockchain built by fra with modular and fra scaling in mind fractal scaling is based on the mandal BR set of fractal images and it explains how l2s become l3s and the entire app chain life cycle goes through its Evolution the frax ecosystem is made of some of the top builders in the space and has stemmed from its OG days of the frax stable coin it is now in the LST space the LRT space B and this frax L2 joins

the existing Suite of fra products including Frack Swap and frax lend you can see frax fairy the bridge to go from any chain arbitrum optimism or main net into the fraxel main net Bridge your tokens and bring your FRA assets into the fractal ecosystem join today to get the best out of fraxel you are still early join today take part in our Expedition into the singularity yeah now the reality is that it's not the only way to do it it is a very natural way to do it but um all all these Technologies one of the things that you learn as an investor is that technology is super

path dependent there there's not only one way that history can play out or that technology can play out you know did it have to be VHS versus betax uh did it have to be you know DVD versus Blu-ray versus HD DVD um a lot of times it really depends on the right founder coming along and just fighting the good fight for a certain thing to win um because certainly possible that the way that it wins is that you know Microsoft creates some Consortium and everybody opts in and you know it's like DRM or something where uh everybody agrees to like this closed standard that is is you know gatee kept by some big companies um that's another way that it could it could play out or maybe it could be orchestrated by the government itself that's another way it could play out um so none of these things are for certain

uh it really depends on the right team the right time the right moment uh to for all these things to come together in just the right way I I think with something like this um usually the rule of thumb and technology is that the solution doesn't get built until the problem is already like horrendously painful you know so it's not like oh yeah I can see that coming therefore let's all adopt this new standard and like start integrating this stuff into browsers and uh everybody start um uh you know build this water marking standard into your image format uh no no no no no people are going to do this when it's already too late uh when it's already oh my God social media is unusable have you guys had the chance to invest in anything similar to this kind of like content

verification not not not yet not yet and I think it's it's one of these things that it really does matter whether you start uh too early like it's it's you know if you look at the history of Technology um there were a lot of people who who had a pretty good idea of how the internet was going to play out and you know you saw in thec bubble um you know the people who are building pets.com they were right right that start up end up becoming called chewy much later and it was successful um or you know web van uh they were absolutely right you know instacart Amazon Fresh delivering groceries yes totally it was going to happen uh but they were just way too early and it doesn't matter if you're right if you're early all that matters is that you arrive at the right

time when all the ingredients are laid out for you to cook the meal you're going to cook but if you're like yo you know I'm going to I'm going to create this you know amazing five-part meal but I don't have any of the ingredients then tough [ __ ] you know your meal is going to taste terrible yep being too early taly no no no but being too early is often synonymous with just being flat out wrong right so you have to timing is everything um all right see well as you spent uh as you said about half the time in your uh video and in your paper on AI we've spent about half the pod on AI which I'm excited about but we have a ton of other stuff to cover as well from your predictions list and just broadly uh looking ahead um so I think one

really cool thing actually just real quick before we kind of close the loop on AI robbit if you have anything as well um like the like the way that I'm thinking about this current AI cycle is I'm in agreement I think that the AI agent fatigue happens quicker uh than you do I see if I think it happens within a quarter or two not a year um I think crypto moves too quick uh but then I think Market matures enough such that right now it's at a like a market awareness level of like two out of five or three maybe um and and sophistication of like maybe one out of five I think those both steadily increase this year um and then I think like the big projects coming on like ritual and allora and um you know hyperbolic and

these if these go to launch tokens and if these go to launch mainnet I think these capture and take the AI sector to the next kind of level in terms of legitimacy in terms of a large um large valuation of token large ecosystem real usage and then I think the strongest of this kind of meme category right now survive and also go head-to-head with some of those and but also kind of in a collaborative effort um and so I kind of see like a bit of like a of a of a handoff of the Baton in a way in terms of like where the narrative goes from like this defi summer group of projects to like this kind of like research institutional group of projects if you will um but

what's particularly interesting is that you is that you mentioned that these are kind of operating within the same convergence but going at it in a different way right whether it's AI flows to crypto or crypto flows to AI so it'll be interesting to kind of see which one wins out I you know we know your prediction but I think that that AI agent fatigue happens quicker than we think and I think that some of these infr projects come on um and kind of The Narrative to the next like throughout the end of the psycho quote unquote um and so I'm I'm curious to see how how that plays out um and yeah Rob any closing thoughts on your and on this AI subject before we get into some of the other interesting topics I have a yeah I have a qu like a zoomed out question

that I think will close this out just kind of zooming all the way out like what do you what do you think is our Mo like you know you mentioned that you're you're already good at you know a lot of the things that AI is good at and I think a lot of people are facing very similar phenomenon or pattern right now where like they're finding that AI is increasingly good at what they're good at whether it's driving a truck or it's riding or it's you know deriving chemicals in a in a or reading x-rays or what have you right what do you think in general is our moat I would say our moat is so the first thing I said is capital um so like before the AI explosion happens go get some [ __ ] money uh so that's that's number one um number two I

would say you're welcome you're welcome uh number two I would say um more and more writing is being devalued and speaking is being upv valued so the more that you uh focus on Charisma relationships communication those things are going to matter more and more in a post AI world because those are the things that AI are not going to be able to do as well right so AI obviously you know if you listen to um you open AI advanced voice mode obviously in a Nar sense it is very charismatic right it has this perfect lilting voice it can sing it can you know has this you know sultry whisper it can you know whatever pronounce everything perfectly never says ums or

Oz or whatever uh never says whatever or whatever when it's giving a response and multiple languages I believe like and multiple langu yeah exactly can speak any language anytime it wants to can recite poetry have just pure memory um but it doesn't get invited to parties and you do and and like the the thing to remember is that you know who has all the resources who has all the capital in the world the answer is other humans at least for the foreseeable future and those other humans do not want to hang out with chbt advanced voice mode they want to hang out with you and so your job then is to be somebody who those humans want to hang out with and so that I I think what what ends up being elevated more and more is relationship skills communication skills and this is

the other thing to understand is that okay now you have a Legion of AIS available to you right what are you going to do with that Legion of AIS can you even come up with something to do are you able to actually orchestrate those AIS and manage them which means that the really important skill for human beings to use all the capital that you've acquired because what does that Capital buy you it buys you intelligence once you have a lot of intelligence underneath you what can you do with it the answer is whatever your organizational and management skills and communication skills allow you to do with it so in the same way okay you've got this great you've got this amazing AI at your fingertips right you've got Claude you've got gbt 4 you've got 01 um can you actually tell it what you want

because if you can tell it what you want it will do it but if you cannot explain what you want then it's not you're gonna think that oh this AI is stupid I can't do what I want no no no the the one who is failing to accomplish the task is you not the AI you so you're saying to invest resources into your teams or your organizational ability the the actual people invest resources into their ability to write proper prompts to and for you on your own as well prompting is a narrow way to describe it I would say and here's the thing we're gonna move past prompting because prompting is like a you know prompting in general it's like oh well there's all these like little secret sauce you tell it to think step by step and you know give it three

examples and do this and do that you tell it nicely tell it you're going to pay it 20 bucks if it gives you the right answer all okay this is like prompting techniques right and these are all kind of weird little tricks of the trade uh this will go away eventually we'll just like bake this stuff into the pipeline of how you interact with an LM so you don't have to know any of that um the thing that won't go away is communication skills right no matter how smart the AI is you will still need to tell it like for example you know um like my um a friend of mine was I was watching them interact with AI and they were saying like oh you know I I want to website you know have this AI create a website for for them and they were like

hey you know I want you to create a homepage for me um that says this this and this and the AI you know just spit you know spits out some homepage and they're like ah you I don't really like it um make it uh you know make it make it make it look nicer and I'm like okay that's you don't know how to explain what you want right like you have some idea in your head of what you want but you don't actually know how to describe what you want and to be clear that would be true if you were talking to an engineer If you hired an engineer and you said you know please make this homepage better and you said okay how do you want me to make it better it's like make it nicer the what what what what the what does he mean make it nicer

right so your ability to communicate what you want and just be able to have the theory of mind to understand another person trying to understand your instructions right that is basically just communication skills that's management skills those skills will be true no matter how prompting changes you will always need those skills in order to interact with AIS in order to interface with other intelligences so your ability to cultivate that skill it's the same skill that managers have right what is a manager's job a manager's job is to understand what their reports are doing and to communicate those reports clear goals and instructions and monitor their progress right that's the basic Loop of what a manager does everybody is going to need those skills because we are all going to have many many agents working

for us and now look the parts of a manager that you want to do is the touchy Fey [ __ ] right you want to be like hey let's you know Small Talk how's your kids are you feeling burnt out blah blah blah like you'll never need to ask your agents that but you will need to ask them how's it coming along what's your progress what are your blockers what do you need me to do in order to get you know uh this thing is not quite what I want I want you to do more of this right can you actually do all of that W you can't then you can have the same number of AIS working for you and you're not going to get the same amount of work out of them as somebody who's better at that fundamental set of skills as a as an entrepreneur for six years I

wholeheartedly excited to have an army of AI agents because it's my capital base that I've built up over the last six years and put that to use to just amplify the work cuz humans are great we have a great team here at the rollup of about 10 now nine or 10 fully bootstrap we Rob and I've just done it on our own it's been a it's been a wild four years but I mean taking that to like 20 and having eight of them be agents or something like that would just be I mean it's not gonna be 20 it's not gonna be 20 it's gonna be 200 oh my goodness so Rob to answer your question the mode is speaking Charisma it's it's really communication skills in the in the person to person in interaction where AI can you know can't kind of step in and

intervene and be as good as us there whereas in writing and in art and design and these things those people are kind of cooked that's right the other thing I'd say is discipline like no amount of AIS are going to make you more disciplined no amount of AIS are going to make you less greedy no amount of AI are going to make you more patient um so all of that stuff of like you know cultivating your own virtue as a human being that stuff you are on your own AIS are not going to help you um and if you're impatient if you're greedy if you're not actually uh long-term focused um or if you're just you know not even receptive to feedback you've got too much of an ego um you know you can use an AI anytime you want to like you know

Andy how many times have you said look here's the business plan of the rollup I'm gonna go submit this business plan to Claude and ask it what what do you think are the big things I'm missing that I could be doing better to get an extra 50% of growth in the roll up um how many times have you asked Claude that I've asked Claude that on the rollup about buying properties about how do I get from exz pros and con like it just it never ends and so you have to be able to understand that and get that feedback but you have to also have the humility you have to have right people won't even ask most people won't even ask most people are like look of course I'm doing the right thing like oh yeah I should I want to you know 50 agents to help me build a [ __ ] food

truck or you know whatever whatever is the thing they want to do right and the their ability to actually take that intelligence and use it to introspect on themselves and to really examine their own goals and what they're doing well and what they're doing poorly yeah um it's like you know if you're a manager you need feedback from your reports you need them to actually tell you what you're doing wrong EXA if you're not getting that level of feedback you're just not improving as a as a manager in a similar way that I think money amplifies SL shows people who they really are and shows the world who people are I think agents and AI are going to amplify high quality entrepreneurs and high quality businesses and also show where the there's a clear

distinction between those with high levels of managerial skills and communication skills and also disciplinary beliefs internally and those who don't and so this kind of Divergence is going to play out across many different areas of life and I'm glad to be on the human side with you boys let's let's keep it human this year all right likewise likewise let's keep it human absolutely all right moving from AI um tons of really cool predictions across the board we can kind of go rapid fire style here um but you know we don't have to touch on them all if we get looped in we can just hang out so one of them was around let's start with the growth of evm market share so

getting a bit more into the crypto weeds to start um we are big alt VM uh people here uh we think think that we are excited about Movement we excited about Eclipse we're excited about fuel we're excited about all these different iterations of scaling ethereum horizontally without needing to rely on the evm however with bar chain and manad this year as well as projects like Mega eth there's a clear return to the evm um so we were just talking about Ai and I'm curious uh as to your point uh with regards to the level of um of ease for agents to actually program in the evm being a big reasoning for why this thesis is kind of like a big one of

yours for 2025 so yeah man I mean if I was to say that I think this is wrong because I think alt VMS will have the largest amount of users and tvl and roll and actual activity by the end of the year compared to evm l2s or even evm l1s you would basically make the argument that well I'm wrong because you know agents are going to be writing a lot of the code and they're not going to be writing it in move so maybe elaborate on that for us and uh and yeah would be would be curious to kind of hear your take there yeah so smart contracts are kind of unique because smart contracts are really really dangerous they are not like normal web development they're more like space shuttle Engineering in that

you get one thing wrong and the whole thing [ __ ] blows up and everyone dies so that's why smart contracts are particularly sensitive it's not the place where you want to be innovating right so on agents it's very clear that uh agents are going to be writing a lot of the code in 2025 and you know I don't know if you guys have played around with any of the AI coding tools like cursor or you know uh repet has their own um agent uh stuff if you if you haven't played around with it strongly encourage you at least with replit it's very easy to use go play around with it and go build a couple apps it's like crazy what you can do even as somebody who's completely non-technical um once you start to see this you build your intuition of like oh [ __ ] this stuff is getting really good

and really easy um you realize that more and more of the the code is going to be written by by agents and a big part of the reason why is that smart contract engineering is really expensive paying smart contract Engineers they go for a lot of money uh requires expensive audits requires you know just a lot of engineering hours in order to build this stuff and make it robust um and the number one thing that you see about these about these agents uh these software engineering agents is that they are much better at certain things than at others so for for example they're very very good at web development and they're terrible at systems engineering now why are they why why is it so different the answer is that the web is full of webdb code so there's just tons and tons of web dev

code out there so they've just learned on tons of you know JavaScript and typescript and python servers and all this stuff that goes into building a website um or or or web app but when it comes to systems level programming there's not that much systems programming that's on GitHub and so as a result they've just trained on much less of it and they're so they're much worse at it so you talk to a systems programmer and they're like yeah I don't know why you guys love these LMS they're they're kind of shitty they write terrible code and the webd guys are like they're amazing they're solving everything they're going to replace my job I don't know what I'm going to do um and so I think what you'll see is that the more training data there is the more audits there are to train on the more

you know infrastructure there's already written or the more libraries that there already are for a given uh language that are more robust and more battl tested the better the LMS are going to be in that category and as a result in those domains the cost of engineering and the cost of creating smart contracts is going to plummet like we're going to go to a world where instead of costing you know millions of dollars to write your smart contracts it's going to be like tens of thousands of dollars wow and that kind of deflation in software is not going to be uniformly distributed it's going to happen in some places where it's like okay maybe in the evm it's just Dirt Cheap to write really robust smart contracts and it might be you know move I mean move is a very safe language it's

a beautiful language uh you know we invested in apas uh you know think move is awesome uh but the reality is that there's just not as much training data and if there's not as much training data that means that you can't necessarily depend on the LM to write all the code for you maybe eventually the LMS will get so good that even with a small amount of training data in a language spec they'll be able to write robust code but right now that's not what it looks like right now it looks like they can write really great code in the areas where they have a lot of training data and they can write terrible code in places where they don't have much uh and so what you may see is that if you want to write and move you still need humans if you want to write svm you still need

humans if you write evm the Bots can take care of everything yep and why pay for hum if that happens exactly now if that happens and to be clear this is a prediction I don't know if that's going to happen especially not in 2025 but if that happens then that non-uniformity where in one place it's Dirt Cheap to engineer in another place it's just you know it's it's a good language but it's just crazy expensive to engineer um almost all of the engineering do like it's like why why would you ever um write a smart contract that costs you know millions of dollars as opposed to write the same smart contract that can be you know 20 $30,000 and save all your money to do BD and do other things or we just not even raise the Venture dollars because like why get diluted you know

[ __ ] [ __ ] the V I was gonna say so I hope that you have invested in any evm auditing companies because those those numbers are going down brother the second and third order consequences of this shift that you're predicting are actually deeper than what you than what we all may as an industry actually fathom currently I think that's right I mean to be clear I I think auditing companies are not necessarily going to go away because there are things they can catch that the LMS can't catch but the LMS like if you think about it the audit reports are going to be much thinner right whereas like you want the human eyeballs because you want the people who are the best in the world and the reality is that right now LMS are not the best in the world at anything they don't write the best poetry they

don't solve the best uh you know they're not better than the best programmers they're not better than the best uh ESS U but they're better than 99% of everybody else right so they're better than the non- phds they're better than the you know the nonprofessional SAS they're better than the non-p poets um but they're not better than the best in the world right and in a way they kind of can't be because they're trained on our stuff so they're trying to emulate what they see in the training data which means that almost by definition how are they going to be better than the best poet like they're trying to be the best poet yeah but by definition if you're trying to emulate Robert Frost you're not going to go above Robert Frost you're just gon to be you know you're GNA be emulating him as best as you can

that's the direction of causation once again like it it it has to learn from something what the best is and Tak in that data before it can really Excel and it's never going to Super theoretically unless it has some weird Network effect that it can combine and cross-pollinate ideas from different areas it it would it would not supersede the best of the best because it always has to be trying to attain and learn because it's a machine learning model it has to have input and yeah garbage in garbage out it has to have a better input in order to have a good output so yeah 100% that's a wild thesis though man like I mean you can see the cost for auditing going down you can see the cost for smart contract jobs going down you can see that entire

honestly man I thought this part of the SE this part of the the uh crypto industry has been a bubble for a while I understand the importance of security but and like the the Rarity of developers but like multiple hundreds and hundreds if not millions of dollars for for smart contract Engineers plus Equity auditing firms just throwing out 300K for a simple audit just because they can like it's just been like this sect this part that is just it's been very ripe for for Destruction so we'll see how that pays out with the evm any final thoughts there has and we will uh we will take it to another one of these predictions one one last thing I want to add so when I say all this I I don't say it because I'm bearish on the SPM or on

move um I actually think that what you're going to see in 2 is that everything grows svm to like total amount of contracts total amount of transactions grow on svm I think they grow on move uh and I think they also grow on the evm the claim that I'm making is that actually evm will gain market share and but I think actually in absolute terms everything is going to grow yep and then I think a a logical claim is going to be after that happens once that level of AI proficiency occurs on other execution environments I think evm dominance is going to be perpetually cucked but we will see over time how it kind of plays out I'm excited about the possibilities for user experience and

just overall uh performance that you know move svm and other I agree with you once once the AI are good enough uh that they can just be given a language spec and then write gray code in that language yeah uh then I my claim stops really making sense exactly when that happens but we're but we're definitely not there yet and it doesn't sound like we'll get there anytime within the next year or two another interesting one is token launches I thought this was really cool where there's a shift from the points meta to basically what you described to be as a two-fold way of launching tokens going forward one is it's a crowd sale so we've seen Legion and Ekko uh really take off I mean Ekko being like the dominant player here um I actually think that Kobe and the

co-founder CEO of coin list were going at it on Twitter the the other day about kind of like you know I think ekko's kind of had more success than anybody really thought I mean everybody believes in Kobe but that thing just took off so that's been a really cool trend of 2024 um and then the other uh kind of way that you predict is that companies will just start to go back to what I am thinking about it as is more of like this uh liquidity mining uh kind of method of airdrop where it's like okay we want to hit a set amount of volume set amount of tvl set amount of users a set amount of metrics and we're just going to direct corly incentivized transparently probably linearly probably based on how who which wallets which

kind of clusters of wallets whatever don't even care if it's sibl or not but who got us to this kpis and so I think this is a big shift from what we've seen this year where frankly it's been fun to airdrop farm this year we did well on hyper actually we miss hyper liquid but bought early we did well on some of the on some of like the pudgy penguin airdrop stuff some of like the uh you know stake this asset for for this airdrop stuff and so that trend has clearly like shifted though so yeah man talk us through this and uh you know I think there's a way that teams can do both of these suggested ideas um do a set kpi amount and then also do a crowd sale I'm so curious to hear kind of what

you think about maybe doing both and then why you believe this trend is Shifting yeah I I think the reality is that people are realizing more and more that the points meta is not really working for a lot of projects right if you just if you just look back over the big aird drops of the last year uh of 2024 and how many of them were positive versus negative right if you think back to igen lir or you think back to you know ziki sync or you think back to you know what there were a bunch of them that just like the sentiment of the airr drop was overwhelmingly negative uh layer zero you know where people were just mad everybody was mad uh whereas if you look at something for example like

hyper liquid where the point system is very very clear like what are we incentivizing there's no surprises it's very clear you do the thing we want you know you provide liquidity you do trading volume you get points that's it easy you know a for B exchange I don't really care who you are I don't really care what the um what the sort of Genie coefficient is of the token distribution all I care about is that I got what I wanted and what I wanted was to be the best exchange kpis are very easy to measure and very simple um and I think that's that's that's where we're going is that once your airdrop is no longer actually buying you positive sentiment and actually buying you the thing that you thought you were getting which is like oh I'm getting the community involved but actually the community

[ __ ] hates me and they're all mad because they didn't get exactly what they want then it's like okay why the [ __ ] am I doing this you know like just to just to do what people think I should be doing like no [ __ ] that um I think what the big thing that people are realizing more in more is that the people who are getting instigated and getting really angry about my airdrop are not even my users they don't stick around as soon as the airdrop is over they pack up and they go to the next thing and so if these are not my users why do I even care what they think why am I trying to reward them and why am I getting fixated in this weird Loop of like finding these non-users giving them money having my non- users yell at me

and then they all leave then it's like well what what was I spending my time on so the the the thing that every project fundamentally wants is one to distribute their token because they need some decentralized token holder base and two they want an army they want an army of people who are going to carry about them and prti for the thing that they're building and um or three they want users but if you're you know if you're like a layer one or you're you know I don't know some like convoluted infrastructure project or you're like I don't know an IP protocol or something uh and it's like well I you know I don't really I don't really know who my users are yet I can't really tell who my users are going to be and I can't really give them my token because they haven't used my

product yet they can't use my product yet but I need tokens distributed on day one because if you're in L1 you kind of have to have your tokens out there so how do you get your tokens out I think the easiest answer now is a crowd sale because it creates that Army it creates that alignment allows people to come in early and it just obviates the idea that like look these random you know farming Rings um why why why prioritize them over people who have some skin in the game and actually are willing to make a bet on me so I think that's where we're moving to is that if you have clear kpis if you actually know what you want then just get what you want just pay for service you know and just like see it as purely transactional make it a linear

airdrop done um but if you really don't have a clear cap you're a lay one you're a layer two and really what you want is an army you just you know you're airdropping for Vibes skip the airdrop and just do the token sale uh and get get people on your side and allow them to participate in in the upside yep much more transparent both ways exactly um and make and making yeah the the token sale as well as the points and you know it might have been a necessity of the bare Market to to you know scoop up whatever users are there but every and everyone was kind of going for the same bucket but um it was also regulation thing is that in the Biden era nobody wanted to touch crowd sales because they thought okay this thing looks way too much like us so awesome that yes so great yes

totally yeah maybe I I have one like broader thing that I I would love to get to be before we wrap up but maybe that do you want to say a few words on like regulation where you see that going because there was a a category in your predictions about kind of like new political regime and how that might change things so what is and about stable coin specifically think is like the key point there from from your talking just one thing real quick before that Rob and then and then yeah we can we can kind of get to your last topic and wrap up here is uh one kind of uh side effect that's in a similar Trend to what we just discussed but kind of orthogonal is the fact that price Discovery is also happening on chain now so that is super freaking interesting

because what we've had for the last so long is that we've basically had tokens Just Launch on exchanges price Discovery happens behind closed doors I'm not going to you know call out VC's right now as we're on a podcast with h but there's this little you know Market maker V everyone has like is on the phone talking about who where the price is gonna launch and how they want to you know or I don't know we're we're not on the phone with anybody talking about where the price fine but I'm I'm just saying there's it is opaque offchain price Discovery is now shifting to onchain price Discovery with this also crowd sale mechanism and these transparent airdrop kpis I think those three things combined make crypto crypto again and that to me is in very

inspiring exciting and just invigorating it's just like okay wow like we made it anyways because of the of the regulation so I'm curious to kind of hear what hasib thinks about kind of that point as well as stable coins and you know if I'm calling you out then tell me why tell me what's actually [Laughter] happening um as a VC um you know when I'm working with a Founder that wants to take their token and and you know go to tge and then eventually you know ideally the project gets listed on on an exchange um like where I get off the train is when we're not doing the product anymore right so I can potentially you know tell somebody hey you know go talk to coinbase go talk to

binance go talk to you know blah blah blah blah or you know it's more often the foundations that are getting in touch with these people and um their the listing process is very secretive and when I say very secretive I mean secretive between the exchange and the issuer nobody else has brought in um and there's obviously a lot of reasons for that because you know these things are there there's a lot of room for for malicious behavior if anybody else is party to those conversations so we intentionally do not want to be a part of that we don't want to be listening we don't want to know about it um we we we can give people the general advice on like hey here's what you should look out for here's what's Market here's how you should negotiate with exchanges um but we're not going to be we're not going to

be a party to the conversation we're definitely not going to be telling them where they should list the exchanges overwhelmingly are the primary party that tell people where to list uh or like what you know what the price at which things come out at and that usually happens between an exchange Market maker and issuer um but no VC's pretty much never have a say and if VC has a say something really [ __ ] shady is going on um like I I can tell you almost every project you can think of that's launching on a tier one exchange uh the VCS are they're not in the room um so yeah you guys get thrown under the bus a lot though so yeah which is fine you know like whatever there there it's it's part of being like public and out there is that you know we have very punchable faces

and it's like cool you know I'll take it um so uh it's fine I you know we can we can roll with it uh I'd say the um the second thing like you know here here's what I'll say if um if random people on crypto Twitter think I'm an [ __ ] I'm okay with that you got to have a tough skin to survive in crypto if entrepreneurs think I'm an [ __ ] then I I got a problem and I need to fix it so that's that to me that's the Northstar is what are the people who work with me or the people who are potentially want to work with me how how do they think I'm doing um as long as they're happy with me I feel good but you can't please people all the time especially not on crypto you're always gonna be pissing off somebody so um

that's just kind of the name of the game yeah that's um that's the signal everything else is just noise coming coming from everyone else out exactly yeah exactly okay so and then yeah any contrarian takes stable coins and regulation uh in in the and that and that onchain price Discovery kind of like tied in there but I think you I'll addressed it yeah yeah so with with onch price Discovery um I I'm I'm a little more circumspect about onchain price Discovery than you are um so with onchain price Discovery two things one is that we've seen this before when did we see it before the answer you said it earlier defi summer right defi summer everything was launching on chain there

were these idos and Unis swap launches and that was the first place that everything listed and you really wanted the alpha you had to go buy it on chain right and um I I think these are usually a sign of not necessarily like oh my God now everything's being democratized it's usually a sign of like oh there's a new speculative frenzy and like the game is you got to be the first and to be the first means you got to be on chain to go grab it early before it gets listed right uh and that being said like you know folks like yourselves you have the alpha and you're advising a lot of your listeners to like go get the alpha which is go buy it before it gets listed um but I don't think it's a okay great now everything's democratized and these things are never get listed you're

buying it early because you want it to list on binance you want it to list on coinbase and when it does it's going to rip right because it's GNA get access to a new layer of retail that's getting in later than you are so you know when it gets the upit listing it's like yo I told you to get it on chain you know props to your boy so I I I don't think it's yeah I don't think you're hoping that these things never get listed yeah I know exactly I mean it just gives it gives the common folk if you will you know I know you're you know you're in the Ivory Tower of dragonfly but it gives the common folk no that's what I'm saying that's what I'm saying look the common folk you're you're not the common folk my friend you run a [ __ ] podcast

you've got Cy th000 followers you're you're not a common folk right the common folk are the people who are buying on up B those are the common folk up those are the common folk yeah yeah yeah I mean look it is ultimately markets are zero sum crypto is positive some in the world but I mean yeah I think right I think that is a bit that is a bit of a of a reality check look I'm just tempering I'm tempering your stat reality being said look I'm a look I'm a def5 bull I'm a defi bull I love [ __ ] defi i' love to see more price discovery on chain we're investors in a lot of this stuff you know um but I also think that CI clearly has a huge role yeah it is the connective tissue with the old world and where people have most

of their money most people don't live on chain yet and until they do live on chain CI is really important to actually get people to to get their money into the secret system so I don't think we should be um [ __ ] on it I think it's a really important part of what makes this whole system work um but uh you know it's kind of like um um I don't know it's again to give another tortured example it's kind of like AOL if I'm you know dating myself to like way back in the day AOL for many people was the first way they got online uh AOL for those of you are not American AOL is America Online which was this this service that brought a ton of people online back in the 90s and early 2000s and um you know there was kind of kind

of a Walled Garden that a lot of people would live on these AOL chat rooms and they thought AOL was the internet and of course it wasn't there was much more to do outside of just AOL but AOL was really important for getting people online and getting them their first kind of training wheels experience with the internet and that's kind of what exchanges are you know to many people they think exchanges are crypto and you ask them like oh you know do you do you interact with ethereum they're like yeah yeah yeah you know I buy eth on coinbase uh and they're like no that's not ethereum ethereum is like a whole platform um that's okay it that's totally okay it's it's completely fine that people's first experience is with coinbase maybe later their next experience will be coinbase wallet and then maybe eventually they'll like you

know get further and further Upstream into like the craziness that you and I live in um but you you got to let them do it at their own pace yeah not everybody's going to start in the deep end of the pool it's very clear that onchain price Discovery benefits those who are onchain I mean that is a very clear objective reality so have to kind of like temper your but also but importantly importantly and and here's where I totally agree with you is that it lowers the cost to entry yes so you know because there's now onchain price Discovery you have a phantom wallet pay a listing fee you have $100 of Soul as a as an average showe or as a project you can spin up and go and you can have a community so that is the I think that is probably more exciting that's the more

exciting part obviously getting in early and then dumping on people is like that's how markets work unfortunately but that is it's it's actually fortunate because it's a it's a ma mature form of value creation and value pricing however uh yeah I mean the reality is getting an early on onchain dumping on Sexes is kind of like what people want to do but I think it's actually launching projects and then having anybody with a bit of Soul or a bit of eth can participate early like you know like you would on on your Ventures arm or like I would on our Ventures arm for a project that is like you know raising it's almost similar for a almost a similar Level Playing Field however there is cabals there is onchain Kabal times it's worse than the same

thing as exchanges right like all that [ __ ] all that [ __ ] being on chain are hum doesn't solve any of that solve any of that it just makes it more visible so the last the last point that I'll make and I I'll wrap with this I didn't get to all your questions but U whatever I'll give you one lesson at the end so uh I made the point earlier that I thought AI was going to be deflationary right and that it was going to massively lower the costs such that people won't need to raise millions of dollars in order to go write some smart contract and deployed on chain and in the same way you know one things that dexes do is they make it so people don't have to pay you know hundreds thousand dollars and go you know do what all this [ __ ] to

get listed on binance or coinbase in order to get their token traded and to get a holder base you can just go on Unis swap you can go on radium and boom even if 20 people want to buy your token you have 20 holders now you know and that wasn't possible in a pre-exchange world or sorry in a pre- defi world so th those things are real improvements now one thing that you might think given my description is that okay well doesn't that put a VC out of a job if you don't need to go raise money to get listed and you don't need to go raise money to write more contracts and you don't need to raise money to do you know all this stuff just gets cheaper and cheaper and cheaper then doesn't that mean that VC is okay well it's kind of done now it's

just you know it's going to be retail investors and people just you know R you've got your C your AI is GNA do for you and why do you need VC's and I I think to me the answer is clearly no clearly no um now why do I say that uh is because think about the internet think about think actually the business you guys run the rollup right uh you guys have a website you guys have podcast you guys you know all the stuff you guys run your own Community um if you if you rewind the clock back to let's say early 2000s right uh in the early 2000s if you wanted to launch a website you needed to buy your own server and you needed to R you like have your server at home which you know very

expensive investment to buy a server and you run your server and you manage it and you do the upkeep yourself you probably have to go hire assist admin and let them SSH into your server so they can run it because it's very difficult to administrate a server at that point in time uh and like that's the startup cost to just run the kind of business you guys are running if you just want like a normal [ __ ] website and a chat room you have to buy $20,000 of hardware and like pay assist admin to run for you right and with the Advent of cloud computing and you know infrastructure as a service and platform as a service and all the stuff getting automated and no code and Discord and blah blah blah all the this stuff now you know what does you guys need to get

started probably like you know hundred bucks in order to start your company right like that's all it really took to just have a website a Twitter account a Discord like all this stuff that used to cost $20,000 startup cost now cost almost nothing right now you might ask the question of okay well doesn't this mean that VCS are out of a job like no the answer is [ __ ] no of course not VC's still have jobs because there's still going to be really difficult moonshot projects that are capital intensive but the capital now doesn't get on stupid [ __ ] that's really easy now right it's like nobody is going to pay you like you would not want to sell Equity because oh we need to buy a server and pay assist admin to run a server for the rollup right like imagine you had to sell 20% of your Equity to do

that [ __ ] no that's a worse world right it's a better world if you're like look I want to scale the rollup up and build a much bigger media Empire and do all this other stuff and for that I do need Capital uh that is where a finance here a VC whatever makes sense and there will always be ambitious projects that want to do something bigger than just the thing that's been deflated down because it's gotten cheaper right so yeah I I'm very happy that as a VC my money is not gonna be going towards you know paying millions of dollars for you know [ __ ] smart contract audits all this other stuff smart contract yeah exactly like all this stuff the the cost of all the stuff will go down great I didn't want to be spending my money on that yeah um

but you know there will still be projects that are capital intensive and require big moonshot things and of course there are some things that the spending is zero sum right if you want to buy a 10 attention the money the cost of attention is not going to go down if anything it might actually go up um and so these kinds of things that you're going to be paying money for the opportunity cost of other people they're not going to change uh and so those things are where your money will be going and not towards the things that are going to be collapsing in cost so it's not as though post cloud computing startups are now incredibly cheap and nobody pays any money to anything and every you know you can build companies for you know 100 bucks you you can you can launch the start the the the

software stuff for bucks but the rest of it still cost money and the same thing is going to be true in a post AI world for crypto startups W well we'll let you know if weever if weever raising we need like 20,000 AI agents so I'm not sure if we can afford that right now all right pretty good pretty good that that that that'll probably not that'll probably be affordable for you guys by the time it's uh it's ready I believe in you all right well thanks for thanks for coming on um and uh yeah it's been a fun conversation my pleasure guys absolutely thanks to