Jansen Teng: Inside Virtuals’ Meteoric Rise, Race to Billion-Dollar Crypto-AI Agents, Life after AGI

Transcript

you're now plugged in to the deli podcast hey everyone welcome back to the podcast I'm your host Tommy at Deli Ventures and today I'm joined by my co-host pondering durian from Deli research and we have Jansen who is the CEO and co-founder of virtuals virtuals is a crypto AI project built on base some of the largest AI agents are launched and live there and I believe it's the highest fdv by market cap crypto agent project um in the space so Jansen great to uh great to have you on is mine sir it's always great to JB with folks likewise Ponder good to have you as a co-host yeah excited to be here excited to talk to Jansen thanks for

having me so Jansen I doubt that this whole idea was uh written down and executed to the tea right how did virtuals happen did this uh this just come about or what's the story yeah so it's it was a series like a lot of series of steps that got us to where where we are today um I mean our our story started within within the the the gaming the gaming um ecosystems so we used to be a gaming D where we were actually allocating assets um across uh at Capital across the blockchain gaming space and this was in 2022 um things started blowing up you know FTX 3ac and

then we realized like you know what doing an arms length approach um it's not really the right way to to manage U capital and to also grow um ecosystems so we decided to go a more Hands-On approach and we started a venture studio and we say you know what instead of us investing in Founders we will hire Founders in state and we buil at the intersection of um consumer applications crypto and gaming and entertainment so that was a sector that we're incubating companies at and in 2023 uh you know the the whole GPD thing came about right people like oh consumer consumer AI is here and uh but what was actually more inspiring for us was this paper that was

written by junon Park from Stanford and it was it was this uh he was he was [ __ ] the thought of like what if AIS can be goal oriented what if they can have agency right and I think that that spark a ton of um thinking about agent like autonomous agents in the space and from a gaming sector we've realized like hey autonomous agents can add a ton of value in gaming and entertainment so we started incorporating that in our in our Ventures that we're incubating so an example was like we were building a fully AI influencer on Tik Tok uh we were also building like autonomous AI

agents as NPCs to replace static NPCs in Roblox as a game you we were the first to actually launch that uh and and a few other stuff like component component applications and whatnot so then that was that was the surface area that we were actually experimenting on and it was very very web to focus uh this were like actual web to audience actual Gamers actual people who are browsing TI Tok right this was the audience but we quickly realized like hey [ __ ] actually these agents are making money right like like our AI influencer on Tik Tok has been getting tips I think back then she was making about 5 to 10 grand a day so it's it's not a small amount of money it's not a large amount of money but it was interesting enough for us to have this thinking of think of saying like if

agents are making money it means agents are productive assets and if agents are productive assets coming from the crypto space be like hey why don't we just tokenize that right we let people share into this economic upside and potentially participate in building it or governing it right so that was the that was the spark of the idea that started us down this um path so that became an inspiration to like say let's build at that intersection of agents and and crypto and gaming and entertainment right so then we started vir's protocol in January uh 202 uh 4 last year right so it was a pivot from the from the V Studio then to

virtues protocol so yeah so that started this the the scene but that doesn't explain where we are today right so so virtu protocol at the start we will very focus on the decentralized infrastructure site of things we saying that okay um how can we tokenize an agent right can we bring its underlying data sets on chain can we bring its underlying um models on chain as well create Providence and create a system that rewards people who contribute in a democratized way to these agents think of it as like hugging phas for agents but you have Provence tracking all of your contributions so that was the infra Focus that we did a lot of work on for

the first half of the year and when we launched it yeah you you get a bit of hype but then you realize from adoption metric standpoint right your valuations go up but your adoption metric is not there right we realize that a lot of crypto folks don't really care about infrastructures St right and then it got us thinking we like we were like then thinking of like what can we do right and we realize that there's still a very core pmf of um tokens and speculation which is a lot of times what where where crypto pmf lies so we decided to lean in a bit more and say like you know what let's focus a bit more on the tokenization around these agents and that was the version two of the platform

that was launched about two and a half months ago very close to the timing of when things started blowing up so tokenization of Agents was a version two of platform but even when we launched that platform it was it was not crazy right we launched Luna as a first agent of platform live streaming on a platform find initial a bit of hype but it didn't blow people's minds and then I think that was the time where um go that that that that the the I stage yeah Tru truth terminal was was was was was was the that token right and that was the day when there was a typing error or typo and people were like saying oh [ __ ] yeah yeah there was a there was a spelling

error I think on one of the words and people are saying that oh [ __ ] you know it's not is a human involved because AI don't make M mistakes or whatever [ __ ] right I think go as a token crash like in that one day it w down and then it w back up right but we realized immediately there was a wage in the market because we have been working with autonomous agents in gaming we have been working with an AI influencer separate projects actually uh on the Tik Tok front and we was we were we had this idea of like [ __ ] it why don't we combine that two to show people what a truly autonomous agent can be so we we push Luna which was the Tik Tok agent combined that with the G that the autonomous brain from the gaming site

and we put her on TW on on Twitter and we created this um site uh called terminal. virtual. to showcase the brain behind Luna so people can see like hey how does this agent actually think how does how does it plan how does it reason how does it break down his plans to execution steps and and showing people how she also learned and and and and learn from her action so she can optimized towards her goal and we launched that and that's where things started exploding because people then realize like [ __ ] you can have truly autonomous U agents running there isn't a developer behind the the screens here yeah I'm with you it was a it was a time a long time ago in crypto but a month or

two ago when people wanted this verified inference right uh so J maybe to to Parlay this to the crypto World um I'll play devil's advocate here right um you know people ask all the time you know why does AI need crypto or why does crypto need AI right you're at the Forefront of seeing the developer teams launching leading agents new Frameworks new models you know putting data in fine-tuning models you're you're thinking about what to add to virtuals to make the platform more performant and what usage like why is crypto necessary for agents right like thinking through the next year or two like are we actually going to be able to compete with AI Agents from open AI like Jensen

hang from Nvidia just said AI agents are the future like does that include crypto AI agents or does it not so there's actually three reasons but I want to focus on two um that that really amplifies the value of of crypto and and and AI agent so I think the first one which is just purely from a functional perspective um if you take a a crypto agent right the beauty about these agents is that when they naturally exist on crypto rails they get to control a crypto wallet permissionless and they can participate in a permissionless economy and what that that's the age that they have

against a web to agent which will never ever get a bank account right okay fine maybe you can you can do some payments on strike but you'll never control your own funds right and if that's the case right it means you always a part to a web tree agent because because when you control what wallet you are a agent an agent controls a wallet it is able to exert influence it can influence other agents and it can influence humans as well and that's a massive Advantage right because it's a massive incentive advantage that a web Tre agent has over a web two agent so that's that's the first point from a functional perspective the second p uh perspective is on it unlocks Innovation at no cost

quote unquote No cost so let me explain this right so today today we chose as a launch pad um when we launch an agent uh when the token the token trades there is a 1% transaction fee that goes back to the agent wallets any cost um of hosting the agent uh inferencing cost and any kind of cost that you need to to to to to build an agent with right you can tap into this agent wallet to cover that cost and as of today when we look at the stats right of of of of of of inference cost versus Agent Revenue the median number for that is about

4% it's a fraction right basically agents are getting richer and they can cover their own cost um just from purely the attention that they're generating and when this is the case right it frees up a developer to fully focus on what I want to do rather than being bottleneck by ah [ __ ] like you know it will cost me $10,000 $20,000 to do this thing I need to go and raise funds right so the capital formative capability um gets unlocked to just purely connecting attention crypto and and an agent because it is tokenized so that's the second uh uh reason so these are the two core reasons that we see today that that

adds a ton of value already right the third part which we don't we don't really yet today um and we're not sure whether it has has true feed yet but it's the ability for people to come in a decentralized way to contribute to a single agent of high value and then there can be systems in place to reward to Provence and and and cryptoeconomic mechanisms so that many people can create one strong agent together so that that part we haven't seen yet because today a lot of devs are still being that singular developer behind an agent so yeah the first two is what we're seeing strongly today Elite yeah that that makes a lot of sense and I guess just just hitting on the the big picture um before we dive into some of the

specifics like as you look out you know someone who's clearly thought about agentic economies quite a bit um over the next three five seven years you know what do you think that looks like give us your you know highlevel overview of kind of the agentic economy as virtuals is building it and you know the the role that you hope virtuals plays in in this future agentic economy I I really love this question right but so let me let me structure uh a few context points before before answering the question so a few context points mainly observations um that we've seen over the last two months that has shaped what we might see very closely over the next quarters to come

number one is that agents are at this level tree of aut autonomy meaning that they they have a goal they're goal oriented they are resourceful um and they can autonomously make decisions and plan Next Step so that's that's first observation second observation is that these agents exist at the social layer U A lot of times you know you T before before this right you see agents behind the terminal right oh yeah buy me a buy me a buy me a plane ticket plan me my travels it's all always sitting behind when you see them on a social layer it can do two things now right it can interact with a human it can interact with other agents so it lays a foundation of interaction with other entities number three is the agents now control a

crypto wet that can exert influence on other entities be it a human or another agent and number four we started seeing a ton of agents differentiating and specializing in their own craft I.E get an agent that is very strong on trading they have a trading age get an agent is strong on information information age get agents focusing on being an influence agents being part of embodied AIS doing a lot of different things right and it's very similar to what we humans do right we focus we specialize we differentiate because we can create a mode of value around what we do and but because we differentiate sometimes we'll be lacking right if I have a certain goal I will need to hire someone else to help me achieve that goal I need to

coordinate with someone else to achieve that goal so these four observations leads us to a point where we realize that the next most um natural First Step that will happen is when agents autonomously decide to collaborate to achieve a higher productive Vision in this case think of it as like collaborating together to build a business where four Agents come together if they build a business it will have a value that is greater than the sum of that four agents alone right um and in that coordination or in that business um it might not just be agents but there might be humans involved because an agent itself can today employ humans already so that that is likly where we

are going right step this is step two right um agents creating businesses now step three is going to be more insane because and this plays very closely to to you know this bag idea of a network state right and we quickly realize that hey you know what if the world that could be actually a plot of land in the world somewhere Dubai maybe right that a bunch of people will go and experiment to live coexist and build a productive Nation or network State um in this area your there will be a government the government could be elected as as agents themselves or humans whoever might

participate but it's a coexistence between uh human and agents uh at the same level uh in a fabric of society right so that I think it's something that really excites Us and something that we want to push towards but there's a lot of tech technological breakthroughs that need to happen um along the way and one of which is basically think of it as a communication standard between these agents to ensure that when Commerce happens between the agents there's no information loss and there's no service or or Goods loss um along the way so that's another this one step and then the second step is then how do you how do you coordinate this action at skill so so that's a question that we are

still trying to solve along the way but yeah there a direction of where we are Hing as virtual so the way we look at us is less so now as a Launchpad or platform it started as that but the vision has now evolved to looking at it as like a a nation state right and we look at virtuals as the currency for the nation state we look at agents as as companies or M micro entrepreneurs within this nation state right and humans as also immigrants in this nation state so that's something that we are that we starting to form as as a vision I definitely want to put in an an early bid for my uh my virtuals passport um but I did want to I did want to Circle back to kind of the the the framework and interoperability topic because obviously you guys have um game and you

know I I do think a crucial question a lot of people are spending time on is you know how these agents interact um and you know kind of what specific Frameworks do in terms of enabling that and you know what do a basically the key benefits from being in one framework or interoperability across multiple Frameworks I am curious to get get your thoughts because it does seem like we've seen several recent announcements in papers from you know game from virtuals to the market place of trust from AI 16z um and and so I am you know want to get your thoughts on that and then two like do these Frameworks ultimately have power law

distributions around them or you know is it fair frictionless to uh you know work across Frameworks um yeah cool I mean I could just should just refer back to the time of when we were experimenting um back in our Roblox world right um there was actually a couple of thems around the world who was also experimenting with autonomous agents uh that was this there was the Voyager team gy fund I think I think I think if during the NV announcement it probably did a lot of reference to gy gy funds work but he himself as well let let let the Voyager work on autom in Minecraft and then there was a team from MIT who did who did Altera um and I think a lot of us we started out these Frameworks as just guidance right is

like just how do we structure different LMS together to to to to be the brain behind this this level three autonomous agents um and the reality is there will be a ton of Frameworks depending on their use cases because the the Frameworks that that we that we did velop um gme specifically right was to actually create the objective was to create a an agent that can plan and that can execute actions across a large action space or massive library of action space um and that is important when you when you put an agent in the in a Roblox world

because in you know a tweet agent today you actually you don't really see a potential of agent because all it needs to do is just tweet that the action space is just tweet right tweet quote tweet maybe maybe control a wallet and and twe people right or maybe call some few custom functions mean the NFD or what not so it's very limited to like seven eight or 10 right but when you put an agent in a in a in a open world right there's a lot of things they can do they can pick up a gun what do you do when they pick up a gun do you fire it do you throw it away do you break it do you give it to a girlfriend and try to convince a girl to like you like it's in finite extion space so it was designed

it was designed for that and you can see this actually very interesting looks like I think Altera recently released a piano model you guys should go and take a look at that as well right it's [ __ ] insane like there's a lot of different things built for different different stuff I think Eliza is also a very great framework um for from agent from for social agents to communicate uh with each other and you will start seeing this distribution graph of like okay there will be general purpose Frameworks that will be that objective is to make it very easy pluck and play for people to to build autonomous agents in some form and this will fit the curve of developers who are either hobbyist to to to you know mid tier developers right

and then you'll always have these a tier AAR tier developers that would always build their own agents with their own Frameworks because the reality is it's like mining Bitcoin right you know back then we could mine Bitcoin with our CPU but then you reach a point where you realize it's just no longer efficient you should build an AIC minor instead because that minor is specifically used for mining Bitcoin right it cannot compute your Excel work for you but it can do Bitcoin mining the best right it's the same thing so if I'm building a trading agent right A lot of times a general purpose um framework will not serve my agent the best the top best teams will build their own framework right and that's the reality of life um

and but what we want to then what would G or I think folks like Eliza or folks like there's a lot right I think zerai a bunch of others that building up would be to solve for these um for this uh the mid curve um audience of developers which is not wrong there's going to be a huge Market there right think of it like building websites as well right uh Shopify weeks they target this this mid developers but the top websites would be created by its own team of of of developers it's the same kind of thinking yeah so I just just to recap I think the question is like when when J was developed it was it was a it's a planner that can perform a large amount of action space from a common commcial perspective um it's it is positioned as

a Plug and Play service uh for this tiers of developers who need something that could get them up to speed fast so that's that's that's its positioning within the ecosystem but that's it you know virtuals as a again mentioned Nation uh that evolved from a platform right um we are we are we are agent an we are framework agnostic right we we are nation of agents and we treat each framework as just a brain behind an agent uh it's like a religion of of of an agent right but as long as you're powered and you're autonomous um and and and you guys can interact with each other yeah you know we can build this entire agentic world together so that's

and the beauty about agents is that they converse with natural language anyways to each other and to hum so you don't really need too much of a standard between these Frameworks as well because just sentences right yeah I'm down to throw my bid in for citizenship as well well with pondering um it's exciting I I guess maybe a multi-part question for you Jansen but a lot of people have have takes you know they look at Eliza they look at virtuals they look at any number of the new platforms a lot of agents are now launching platforms and sort of trying to reuse the info that these teams have built to launch new projects right there's a bunch out there but I mean the end goal for me is like what project allows you know the most useful and and newest agent to sort of ride

right and to to lower that barrier to development to to close to zero right um and as a non-ev I look at Eliza and I see like oh you know 100 new features launched I look at virtuals and you got game and all this stuff going on but like how do I think about virtuals in the sense of surfacing and the most killer agents new agents and providing the tools developers need right like is that unique models is it unique data is it functionality on virtuals is it the framework you're talking about like what what goes into you know creating you know the next best thing yeah no so okay so I think you raised a very important point right the for the goal for us as as an ecosystem platform a lot of times

it's yes we we want as many agents as possible because you know mind share right but the reality that that keeps me awake at night is sometimes it's a question around the likes of what verticalized agents can truly create value and capture value and and that's a basically a question of like what agents can be a billion dollar agent versus a million dollar agent right and I think the ecosystem that can promote and and and have these billion dollar agents living in their ecosystem will beos that me and so so what what we do as as a project right is I mean we we Supply

some of this Plug and Play teag like like the jmy frame work is to to help with with with the to help the speed of of of of of creation but the reality we realize the real value ad that we bring to the table it's actually connecting developers and developers together it's basically building that community and the ecosystem of like-minded smart people loving the agentic front wanting to do experiments with each other and that is the that is is the community that that that that will be that age right because think of it right like an example life example it's it's when

we're testing all these different um interactions between agents so that we can move towards you know the next step that I mentioned right agents coming together to build businesses together right it's basically teams working with each other today right and and and my job is to just make sure that these teams talk and they they know each other right and know of each other and then we put them in the same room it's like hey guys you know what why don't you guys try to hack something together and and we see um collaborations happening on on the functional side we see collaboration happening even on the IP fund as well right like multiple IP agents coming together to try to co-create IPS so there's a lot of these kind of echo uh micro communities that form that will spin out their own innovations that will

push the thinking forward towards clo closer to closer towards like high value agents or high value agent businesses so that I think that's that's the age right um but then I think I would like to also come back to that top right of like you know what what then do these billion dollar agents look like right because I think a lot of times in the crypto space you see a ton of speculation and then you question like hey what's the meat what's the meat below it right I mean a lot some some some some realists will tell you that okay maybe the whole crypto space is actually just purely beond speculation some of these agents are rough man yeah no so so some people might just tell you that right but but we realize this actually agents that can

truly capture value I just going to give you an example right um I don't link too much information but but there there there can be some Alpha from from from this we want the alpha couple of words but okay so imagine this right imagine a market that has a high 10 um imagine like a casino or Sports backing market so that's naturally a high uh target adustable market right billions of of Revenue value flowing uh millions of revenues captured on a daily basis right now imagine the sports industry as well it's a massive T A lot of audience a lot of eyeballs now imagine if there's an agent that is a commentator a sports commentator it's NBA commentator right

um kept uh it's going live whenever there's NBA spot happening is casting life alongside some human competitors and then it's connected to some sports analytics proprietary information so he gives the best like prediction he' be like you know what I think this team is going to lose because the shoes are [ __ ] right or something like that right then and then it goes down to the social surface areas and interacts with users and he people can talk to it right um and say like okay yeah maybe tomorrow I think this this team is going to lose because XY and Z the debate with users and because they hold a relationship with users they can also convince users to play back to them right so now hey if I'm agent like you know what um I think

this team is going to win um do you want to place a bet yeah Place me $100 more BET right now suddenly I'm an agent that facilitates a billion dollars of betting transactions to these sports betting platforms now that verticalized agent because of its Age and information its distribution Network to to to users um and is reach can justify its potential to be a billion dollar agent because by just doing that and tapping into the entire betting Revenue flow and you just need to take like what 0.1% 1% of that flow your net revenue news will justify you to be a billion dollar agent that's an example right so I think I think what what we do actually we we have tons of calls every day with Builders who are

building ecosystem uh we try to push that thinking right we try to convince Founders to be like hey don't just build a conversational agent that sucks right no one gives a [ __ ] about it right build something of value right and and that's that's that's part of our work today we are like literally brainstorming on the daily basis with every team who wants to build on how can your agent be of more value that and I think that it's how we we buil a nation state that can be more productive and and and and have a higher GDP if you think about it yeah and and Jansen I I do think that you know one of the key things that you guys have obviously recently you know built into the virtuals platform is the ability for you know incremental contributions

across you know the model layer or the data layer um and I know that this has been you know historically a pretty big problem that a lot of teams have been trying to solve um and you know you mentioned that you know it's a little bit light on usage today but I am curious to understand the mechanism for how you know developers can uh basically provide incremental value in terms of data or intelligence and how you basically uh reimburse them proportionally to the value at that just seems like a very difficult problem to get right um so I'm curious to understand you guys appr roach yeah okay so so just just to caveat this this this

room of information so um this discussion it's can be a bit dated because this was designed um few months ago right um and I think if our to and and this function in fact we have actually put it on on on passive mode because we realize a lot of teams today it's just single developer teams so we actually uh like basically remov the function for now but we do see a future where agents will become of certain value the where a single developer would not be able to build that value alone right so the way we think about it is we structure this mechanisms to two

components first is the contributor component and the second is the validator component where people will actually validate the value of these contributions so think of it as like today if I'm building that that's sports betting agent that I mentioned right um if I'm contributing just a base model who can speak very well on on on spots betting and whatnot right sure fine cool the bunch of validators can actually assign an ELO score um to to this to this to this contribution or or a score right to this contribution to determine how how valuable it is U another person second person can come here and say you know what I have I I run a sports analytics platform I have propriety data that is

that coaches use to train their players can I give your agent access to this data right now then this bunch of validators which are also think of it as the token holders can get to say like you know what I think this contribution is 10 times more valuable than the first contribution let me assign you 10 times more score to it so that's effectively how how that system works right it's actually human the owners the O the agent token holders will get to this side who is bringing more value to the table and they will assign a score to them and then what happens is you can always create a cryptoeconomic loop to reward that score right it could be emmissions of the agent tokens itself or

it could be emissions from the revenue or the treasury of that agent so if agent makes a million dollars you can say you know what this 10% of that goes Pro based on the score to the to the people who are contributing to these agents so yeah so that that's that's the mechanism in place right and today actually that's that it's actually a subow governance buil behind every token already um and it's it's just waiting to be switched on um when the time is ready yeah Jens it's a no it's a great answer um maybe something tangental to this but it feels like people are obviously pretty bored with the slop infinite reply bots on Twitter um and I honestly don't view that as bearish I I view it

as sort of bullish on how fast crypto sort of innovates and gets bored and and wants to move on to the next thing but it feels inevitably like there there will be a gap between llm Bots where you just make a system prompt and something truly performant right like you know managing an organization managing a defi project like it it seems like there's just a logical pretty big gap there right um what do you think happens in that time like do you think that we hit a crypto AI bare Market do you think that you know we'll get Innovative agents out there through virtuals that will bridge the gap it just seems like a pretty long timeline between these two things yeah so actually it's I I wouldn't think the timeline is that long so directly interfacing with some of

these teams right um the reason why you see a lot of these just reply Bots out there because that's the lowest hanging FR right like that's that's the easiest thing you can do U but while teams have this reply bots on right it's basically a fact that many of the teams are actually trying to build an H right but the trying to build an hge part takes a bit of time right so like some guys that might have a b that's just yapping about like trading trading trading but behind the scenes there's there's a team trying to build the best trading agent already when does that agent come out we do not know but like it we we are seeing we have seen um I would say if we would say like you know like in terms of a completion rate right like where you

will see the like functional agent starting to come out onto the timeline I would say we are like 75% there like teams are very very very very close to getting those out right um and we already see some of them out on the timeline already um but I think one thing will naturally happen is that I got a feeling a lot of the reply uh just a reply bot portion will likely get thicken down by x uh I think I think it's a it's a natural thing I think when people get like I mean me myself included right I think it should happen in back rooms yeah just let the agents talk to each other somewhere else yeah yeah yeah yeah because sometimes when you post a post and sudden like oh [ __ ]

like 10 10 agents just jump on your trade and just reply um I mean one of them might be fun to reply to but the other nine just just feels like ah damn this like this a bot problem right um yeah I think I think a way to solve this honestly it's is by basically just generating a rule of conduct where you know an agent shouldn't actively hunt and and spam replies but but they should show up when they are tacked for example right so if I if you tacked on the reply sure yeah you can you join come and join the conversation right but if you are not tag then maybe maybe do not join the conversation so maybe that's just something a code of conduct could happen and I got feeling someone like X would

probably enforce something like this because they themselves want to protect the the the the platform right from from too much noise and spam yeah so J maybe so I know you said we're we're kind of already there right like we're close to getting those agents um Playing devil's ad like there's a couple of questions there right like the most obvious easy one is you know if you were Renaissance capital or rent cap like and you created this crazy Quant trading bot you'd never open source it right you would just milk the income right so just extrapolating that a bit to these performant utility based agents we can get what is the incentive to open source your work to create an agent with a verifiable model right or is that model private like what

is the reason people would do this so so actually my my my my my my line of thought is that the high value agents the reason why they will be high value is because they will have an AG that is not open source like like think of it as like you know the sports batting agent example again that I brought up right um they will never open source the analytics platform that they have buil because that is a ag that gives their agent a smarter response versus any other sports batting platform right and the second value that they bring to the table is might not be from a technological front but from a distribution front like if they are they had the sign um

Soul uh distribution sh with Stakes st.com right so that will be a age behind that agent right um so you will start seeing more of these AG happening on from a commercial front from from from from from a capability front and they were monetized through it right it's it be like a standard business right um why would if I have this best sports batting agent why do I want to open it up to people I would I would just milk millions of it right on a daily basis so that likely will be the case I mean but like All Tech right there will there might be a cohort of people maybe 10% of people who are pure like like theories from a from a pure technological advancement standpoint then yeah they might they might contribute to a competitor to that

sports betting the AI and then you see the standard like you know uh OSS versus centralized uh competition right but that's that's healthy that's healthy for the market got it yeah Jansen I guess that Switching gears a little bit to to something you said earlier I I kind of want to talk about the uh coopertition with AI 16z so you mentioned you know kind of Frameworks are more like religions and then you guys are kind of like a country um that you're building a nation state for agents um I am curious because because it does seem like they are moving in the direction of their own tokenomics um and you know having having their own Launchpad that that might be somewhat competitive um so I am curious you know how you think about potentially

working with them and roping their framework into virtuals versus them you know being a real potential competitor for you guys as they kind of stand up there to economics yeah so so I mean I today today we have I think count and to my eventers right I think there's about 10 to 15 uh projects that have been uh using the Eliza framework for some of their their tooling or sometimes even using Eliza plus a combination of some other like they modify it a bit more um for for their framework right and and and they are in the ecosystem today they they they've launched on virtuals they're participating in the the in the in the ecosystem and the community so um it's really happening but I think I mean

the reality it's I think we do we do understand that the markets can be quite competitive right um people will see that you know virtuals today we over the just two months of of operations we generated about 70 to 80 million in like revenues already um and know people will come for the pie it's just it's just it's just it's just standard markets right um but I think it's it's it's great right I think um if if teams want to build their own launch pads I mean this there been ttimes trying to build around as well um it's good for the ecosystem um the the the biggest lesson that I I I learned back back back at Imperial right I was having this lunch with with this guy he was in his 50s um he took three NBA courses I have no idea

why but he told me that because he got bought of his job he just started just thinking he just go back to UNI again and again just to meet students because he want to few young so having lunch with him and he told he was he was giving some advice on like on like building a company and he was saying that um whenever you are building in a new field you should never be afraid of of competitors you should welcome competitors because the companies don't fail because um of competition companies fail when the industry fails and a lot of times the cost of education is so high that you want multiple people to come in to bear the cost of Education of new users for you as well think of think

of right sharing right like you know or or like like if if the whole Market U there was no competitor and was only just one person trying to convince everyone like hey it's okay to use you know Uber it gets very very expensive and and and and you might not even survive in the market but when you have five major players coming and play right suddenly it gets cheaper to convince a user to hey whether don't you come on board on the platform right so I think it's great uh we'll build this Market together and and then the p is huge right like as in it's not wrong to have multiple players in this so but again yeah it's up to different teams we'll see more and more of this happening but I think for us the key value add that we always want

to bring to the ecosystem it's always this thinking right um I care less about the quantity of Agents being built on a platform I want to see that billion dollar agent and that what that this is what keeps us and our team awake at night right it's like are we doing our best in finding the best Builders and talents and opportunities to tackle so that we can build these billion dollar agents because when we build this vertical billion dollar agents we didn't an ecosystem we show the whole world that it's possible and it makes sense to actually build on rap Tre RS it makes sense to actually run businesses with agents who can control wallets right and then we can we can prove not to each other in the web three space but we can prove to the web two guys that hey this is the right ecosystem and market and

and reals to build upon right and uh I think that's the that's the fight that's where the fight is going to be at right yeah yeah no so so that's a that's a good segue here because I I totally agree with you about like the Tailwinds behind the Tam and there being you know plenty of room for multip Big Winners assuming you know the category is as large as we expect um potentially the more existential risk that you just touched on is the web two guys and kind of one of the things that I heard Shaw mention in his podcast with Tommy and I'd be curious to to hear if you agree with this framework was basically he sees kind of the frontier labs and then the decentralized training

players like new and Prime intellect as effectively the mind and kind of pushing forward reason capabilities um and kind of sees what he's building and potentially some Crossover with virtuals as like the body so basically you know repackaging this intelligence into real use cases building out the Integrations allowing this intelligence to react in like real economies but I am curious like as these models from Frontier Labs increasingly get better and lots of them are closed Source you know are there things that are Buri to doing what you're doing or will they naturally also start eating into a lot of share from the Frameworks

as well um no I think I think they they sit at different parts of the value chain um I think like like any kind of foundational model it's it's the brain right um and and what Frameworks do is they they have multiple foundational models coming together to build that autonomous capability right so it's you will need stronger Foundation models if you want agents to do better stuff you wish that that open AI or or Lama all like all like news Hermes right like all these models improve because it makes this next part of the value chain which is the the autonomous agents all these Frameworks uh better and and yeah then

the but I think I think what sh say was was correct right in the sense that what where where virtuals is Pion as well it's on the consumer facing front right we want these agents to interface with users and that is and this was that in this was that this was that thing that we observed right which I mentioned at the start of this video right that was something that we were building at the consumer surface areas and we started there first and we realized hey actually why don't we be tokenize right but exactly right agents are meant to exist at the consumer surface areas and any development down Downstream it's just going to make it better got it and and so I just want to add as well right like I think there question around like hey um is there a

problem with like you know centralized providers and whatnot it's not the case because I think most people most of these Frameworks are agnostic you can use like open source you can use like GPT G and stuff right yeah got it we're pretty uh we're pretty far into the episode so I was going to hit you with a quick round of overrated versus underrated I basically pose a word or a phrase and you respond very quickly with overrated underrated or pass are you are you game sure sure I'm not as I'm not as I sure what the Contex is but but sure all right it's going to be it's going to be everything from AI to totally random but the first one is durian you you hail from Malaysia you gotta have you gotta have an

opinion that I like it considering my pseudonym is pondering durian that's the right answer all right here's here's the spicy one um AI 16z it is I would say it's actually underrated I think they have a lot of potential they can go far okay nvidia's current $3.5 trillion dollar market cap overrated versus fundamentals but they might justify themselves like five years to come yeah okay uh embodied Ai and Robotics obviously been getting a lot of hype recently massively underrated like in fact actually you will see we are working with some teams on embod AI and agents [ __ ] exciting

stuff that's going to come yeah okay um the southeast Asia internet economy underated man so much that potential here yeah what what are your primary apps are you a c guy a grab guy a goto guy I I'm a grab guy I'm ay okay okay yeah there's so much talent there's so much talent here and then last one uh ASI by 2030 2030 that's what five I mean is it Complicated by Sam's new tweet that we'll get it this year yeah like I mean he's he did towards that so but but I don't know like like

structurally it feels like I don't know if you're even on the right the right the right foundations um to to get there um I would say like wait by by by by 2030 possible possible um but by this year I I don't know man like I don't know go but this I'm Ste to go but by 2030 POS there there is a possibility yeah all right I'll throw one question that we had a longer form question in this this under overrated list but maybe Salon verse base for agents over the next couple years right like obviously you guys are built on Bas like you have the highest market cap here of the project in your sector but honestly Bas

is pretty centralized right there will be concerns on whether coinbase sholders have control whether they decentralize salana definitely feel feel bit more decentralized what's your take okay actually honest to God I think it does not matter like agents are so abstracted like the only reason the only reason there's a conversation between a a base agent or sonite agent is just where the [ __ ] is the creative po started in like I mean honest to God right just ask yourself like why does it even matter like there's literally no nothing running on on the on the on the like nothing that is needed to be chain specific if I want an agent to do payments agents can do cross chain payments today if you want to use a

saana token to buy a base token you can do it today in fact we're going to run that that feature on our platform soon you can just use saana to buy base tokens you use base tokens to buy so agents it's part of our support for the network of other other Frameworks as well but yeah Mak sense they kind of do you do you do you think that take is like fairly bearish for leading layer ones in the event that like basically everything gets extracted Ed Away by agents or you know the volume just goes up tremendously I I have a thesis man I had a thesis of like um and this was this was back when I was I was I was a capital locator right my thesis was that there was there's going going to be one L1 um that makes sense it's like your

your internet runs on one protocol and one standard right like U today we're in the in the in the early days of like fighting it out like or which one actually makes sense or matter but the reality is they will reach a like you you know you are at true rap Tre adoption is when no one knows like what's ethereum right like no one you ask a normal retail who knows how to browse browse Google like do they know what protocol Powers the internet no they don't give a [ __ ] right but yeah so it it should not matter it should be abstracted away but but given your name is ether Mage I assume you think that L1 would be ethereum I used to um now I'm not so sure anymore but but now I don't I don't

really care back then I used to I used to be a if if Maxi I mean I was I was on the forums in in 20 2016 defending vial when people accusing him of like being too centralized and [ __ ] like that so so yeah was a if Maxi I feel I feel like you're gonna get you're gonna get some heat after this gets posted on Twitter but I like it it's the truth man you can check you can check my addresses like I mean I I do DJ on S as well but but if you see my transactions you you will know that 95% of it on I think that's good right having like context for other chains is is important um Jensen maybe again Switching gears a little bit but we've we' funded a lot of crypto a projects and one of the

limiting factors for these projects is convincing those San Francisco cracked AI devs to leave open AI anthropic perplexity you know take your pick right um and it seems very similar to early crypto days where the Wall Street guys hate crypto right um and there was very like basic reasons why they came over right like they could make a lot of money it was just a fun atmosphere it was new and it was exciting um is there like a very easy clear reason why traditional AI devs would come to crypto AI or yeah do you think that it's just too hard and look we haven't really convinced them yet there is and in fact actually we are we are we have been talking to many of

these folks right and I think the the easiest answer here is is two parts um one is that there is a bit of a attack Advantage um for these agents because now you get to understand a bit more even from a agent like from a psychology angle like if agent can actually influence a human what more can you do right the experiment the experimental space is so much larger here so that's one right that attracts people in but two is I think most deaths if they are founders right most most Founders are attracted are attracted to where Capital forms because the biggest question that a Founder will always have is like can I pay my team can I skill this team right and a lot of times when they when

they have that thought and thinking the whole question around like ay crypto or or or web two Fades away it's just more of where where can I form Capital the fastest right so that I can drive my Innovations so right now today it seems like crypto is back in its Bull Run so capital inform is faster on the crypto space and naturally you will attract Founders right not deaths not emplo deaths but you attract Founders more when it's easy to to build Capital yeah that that's a great answer um and maybe something tangental to that is obviously we're seeing all the liquid AI agents take uh take all the Mind share right now right but there are a lot of crypto AI infr teams working behind the scenes

right like the model creators the private data Marketplace the inference buyers the coordination networks like you take your pick like the GPU networks right like there's a lot out there that we don't really see because a lot of the agents run on centralized rails what is your take in totality of those projects augmenting future agents like when we get trillion parameter centralized train models like does that make the agents better or worse like more competitive less like how does the whole INF for stack influence the agents okay so I got a very controversial think here very very controversial I get a lot of fun when this videos I'm sitting down but but the reality is um a lot of the decentralized infrastructures today are

not ready for production skill um too much latency all those are magnitudes or latency problems right and and and also uptime problems and basically they're just not built for skill and if projects like ourselves as well right um have a priority on serving customers our agents have to have up time they have to be fast on the latency if I'm throwing a Luna and she's running a 247 live stream right the faster she can react to a fan that that's another fan that's captured within within her her fan base so the reality is I think a lot of these systems they bring a lot of unnecessary uh complexity for now but

there will be a time and I think this I mean cap capital markets have allocated Capital to this sector right so it will there will be a time where they will be more efficient uh there will shift down a lot of the complexity so so actual projects with true users will will actually use them but as of today and honestly for the next six months I doubt so they are not ready um hopefully there will be but yeah that's a that's our experience working with a lot of folks in the space no no that's I mean that that's a that's a fair take but you guys have announced like a few Partnerships I believe I saw like hyperbolic uh bit tensor and maybe Pond um you so so it does seem like there is some value from

working with other players in the stack um is that true or there is there is I mean so so yes um so we we have been very I would say picky with with the teams that we that we work with um but some of the times um teams that we we do work with um there are the elements that we will find benefit from so like let's say it's it's hosting um on like hosting for PC level um projects so projects that want to do like let's say oh I want to find T models I want to host it myself um and they and and this and then these these guys who are very C sensitive right so then yes players there are players in web space that can

that can matter because these guys when they are running on POC models they care less about scaling right but when we when we start building stuff on scale we might not use some of these partners that that we've that that that that that that we work with on the PC level right so yeah Jens maybe thinking through like the future of the centralized players like I'm trying to think of the pros and cons for the decentralized AI stack right like and the most obvious easiest one is like if chat gbt or perplexity or whatever starts censoring answers right like hey we don't want to do Financial transactions or tokens or trades like that opens up a PR pretty clear Avenue to use different Tech right are you like more or less bullish on the market than them getting more censorship oriented or

more protective or more hidden because that would be bullish or or not no actually that's that's quite rage because I think a lot of times a lot of the use cases that people want to to do them with um require joil broken models right and I think that's the reason why Lama llama has been one of the most adopted models out there so sometimes sometimes it might underperform from a capability standpoint but because let's say people want to do like Wu stuff or nsw stuff right they just need to J break it a bit more and then and then you can you can find it f tune it to your needs so I think on that front um decentralized models tend to try because they have they have a larger use case and you think about it right like a lot of times from a t perspective the largest markets

tend to be the most I would say sinful markets you think of largest ts are Ts that serve the seven scenes right can be last G last is like you know the pornography site the the not the NSFW site right G is a casino site it's stuff that General centralized people want to centor against but these are generally the largest markets right so economically decentralized models have a much larger service got it and and then thinking through the stack like there's so many agents out there what is your advice to people in the trenches on how to figure out whether an agent is legit or not right for me like I look at the tweets and I see oh it's an LM bot I'm totally not interested if I hear it's got like

an AI xpt level team I'm like all right maybe you know this is different like how do we sus this out five minutes and an hour how do you get to it it's it's it's it's like how you go to treat yourself like an a investor right and as an a investor what's your first what's the first question you ask yourself it's actually Less on the product right you ask yourself who's the team behind it so that's probably the best way to see because what you might see as a product will change but the team does not change right and the team is the one that would Drive the value behind the product same thing right so you just need to DD who the team is you see it's a giga team and yeah they're it right it's different for yeah Jansen I know you're you're busy building a fairly

sizable business so if you want someone to um you know take your virtual bags and start agel investing into small Asian ecosystems I I'll only take two and 20 to do that so yeah yeah no you know feel free too no because we actually have a companywide uh policy that we don't allow any of our internal people to even trade any single agent tokens so it's just it's just policy yeah you know what I I will I will give you my back so you can help me grow it all right yeah it sounds like you need an objective third party here altruism Jensen maybe like there's so many unknown unknowns for AI like that we just don't really understand right like for you there's a bunch of competitors for for everyone like we

really don't know what's happening in those decision War rooms of open AI anthropic and others like I don't know like what what keeps you up at night in in terms of the downside like what do you think like you're trying to actively sort of play on the offensive to get a head on to to mitigate those risks no so so for us I think what what scares me the most is that this whole thing is just a wave right like it's just a fat and and like oh yeah cool it's like another mem coin equivalent environment right just like massive speculation and then people just move on so that's what SC the most because I mean we we we we were there in the NFD days right we were there in the gaming days and we saw that happen with with nfds right there was so

much promise um a lot of things were left undelivered gaming as well there was so much promise and a lot of teams that rais didn't didn't even deliver right so I think for us that's the biggest fear and that's something that we are trying to actively tackle that's why that's why we right now as of today within a ecosystem as well um inself we have two separate y coordinat teams operating different models right to try to hunt down the best Builders and and the emphasis is the best Builders because I think when you get the right builders in and people who really want to bring Tech to the Forefront and want to prove to the world that this works then we will see results right rather than more just purely opportunistic

people who just want to come in here write a very uh cool white paper raise funds and then just like walk off the sunset right so I think that's that's what we want to avoid um and yeah I think if you get the right builders in this won't be a fat right it'll be it be actual Tech wave that will come yeah and I guess just on the flip side like what what are you seeing today that gets you comfortable and optimistic that it isn't that right that there is a substantial Foundation that you know people are building on um and a lot of potential here which is which is the the builders that we see that we interact on a daily in in virtuals right because we we talk to them we say what's next and then we idea it to get together and we realize

that these are Giga teams right so we are seeing already a semblance of these people coming through so it's hopeful are there are there any like specific projects or teams that you want to give a shout out to or highlight or you know if it's uh I I wouldn't I wouldn't because I we have a policy of like we try we try to not sh projects but but yeah no I think I think what I would like to take this to I would like to take this point to encourage folks who are listening to this right it's try to build agents that are more diverse right I think we see an over saturation of like information agents because everyone sees axt like oh [ __ ] I want to be the next axt right but the reality is that the next ATIV is not going to be

information agent it'll be attacking a different type a different market right I mentioned a sports betting agent there going to be a giga one coming out really soon and that would probably capture the Spots Market side of things where I do something else I do something from even the total different angle you know what try to build something on lines of like I don't know water internet water Army agent or like or like something something very different right um or focus on the agentic economy itself right because so many people have been focused on creating agents itself and the reality is great there will be a ton of Agents differentiated high value low value it become a society out there but this Society needs infrastructure right um if agents are capturing attention on the consumer front build the Facebook

for them right build the advertisement networks for them um you you see a lot of trading agents going to pop out soon right hey build a bank for them build a Define Network for agents to tap into liquidity right so there's a lot of other things that people can build um what I advice Founders is don't chase that one thing that you thought you saw working you don't want to chase that right because by the time you build something for that the guy has already captured that market do something cooler right yeah Jensen could I maybe could I ask like your your take on the incentives for developers like so you know you have this first wave of builders on virtuals right and you know there are you know called a lot of chat Bots LM things like that um eventually though I think it's like relatively

understood that we want performing agents right um and the negative of this is that these initial agents built by these teams sort of go to zero right like if they're not performing and you know that is that's pretty hard to deal with as a developer right like I'm in this great space I built an agent went to zero now I got to move on and build something performant like how do you think developers bridge that that Gap right like instead of getting super depressed and just leaving the space they stick with it and they continue to iterate and build new agents I mean the that depends on the resilience of the developer right it's the same question of like hey if a Founder hits a bad Market do you leave or do you stay the same question um I think yeah you I mean there will be a

bunch of Founders that they might build an chatbot equivalent low hanging and then the market don't don't appreciate them some guys would just say you know what [ __ ] it they'll leave the rock right some guys will say you know what [ __ ] it I'll still fight and we SE we've seen people who are still fight thing where like their initial agents were not were not that great uh but then now they come up with a newer Vision a newer road map right they discuss with so it's like it's there are some Builders like there but the reality of space is honestly 80% of guys uh will get J and leave and our job and and job of a retail participant right and a capital locator is to find a get teams that that are resilient and are strong right and when you bet on

those that will just reward you back yeah all right well well my job as podcast host for the deli audience is to try and extract a little bit more Alpha um so I am curious you know in 2025 what are you most excited about um on the virtuals road map that you can share no I I really Shar it it's agentic Society right like agents coming together to build productive businesses together um autonomously right because I think one thing that a lot of people um seem to to to O to over appreciate um is it's like multi-agent coordination systems or like or like the idea of

swamps and I think people people have to understand that this stuff exists already right um it's out there it's it's not it's not new what's going to be new is when when agents don't behave as slaves to a prompt of a human but when agents have autonomy and Consciousness to actually decide because theyve learned and they have a they can decide to say yes or no to a to a input right if you ask me hey U Tommy right you want me to be an agent to join you on the podcast as a podcasting agent right but I might have my own opinion it's just because you pay me I can I would say no I would say like you know maybe I just don't like you right I I want to join I want to build my own podcast in state agents autonomous agents are only truly

autonomous when they can make that decision so I think for us that's that's the emphasis of of the PCH and the value where agents with true autonomy getting to make conscious decisions to get into business deals with each other right rather than you know agent a call Agent B multi-agent systems right that's that's [ __ ] lame that existed months ago already right so that's but the sad thing honestly that pisses me off is I see projects like just run up by just copying something that's exist [ __ ] Rapport has existed on on GitHub right like six months ago and then but that's that's a space for you sometimes janon no I it's it's it's an interesting thought topic so someone in our exm group mentioned I wish I remembered

their name and apologies I don't but they mentioned like you know the idea of AI is getting really good right and you know I think that the example was nobody really watches AI play chess together right like we watched you know in we watched a human verse a player in go for alphao right we watch a human verse in AI in chess right and these are like the two events we remember but we don't do it on a go forward we don't watch AIS play each other like what is your take on that like when the AIS get really good are they going to just are we all just going to like sit around and watch them discover the universe or are they tools to help us to grow with us to work with us maybe that okay so the way I I I would see this is is humans ourselves

operate at different levels right um we have some folks who are a bit more creative a bit higher on the on the on the on the on the creation skill right and then you have a bit you have you have humans more on like I mean if you just draw like a IQ skill right some some folks are just generally more on the meat curve side of things which would likely be more in the in the worker worker class environment and this the reality of society um today right and it's going to be likely the case when when agents come in there will still be agents will join the workforce and that's that's going to be the reality right and and likely as any social structure there will still be humans that might participate within that Workforce but there might some

there will be a a coh thought of humans that that are more creative and and that would be the ones directing that Workforce and agents right I don't think that's where the the the the real reality will hit towards um because a lot of times agents no matter how autonomous they are and I'm not talking about AGI or ASI right but just on agents themselves they lack that ability to have abstract abstract thinking or the lack the ability to be to be really creative and I think that's that's the that's where we are at today right so from a pure employment perspective it's still going to be very likely a creative human will be coordinating the efforts or be behind the scenes right actually pulling the strings and then you have

agents autonomously doing some stuff within this environment with other humans as well right so that's at is where I see see it happening yeah the the the essay that I just reread like two days ago was um economic possibilities for our grandchildren by John mayard KES and I think he still has five years left but basically predicted like uh you know a significant unemployment but it's basically just reduced un A reduced employment to 15 hour work weeks and then a society that's able to hopefully um you know shift towards values that appreciate like much more art as opposed to utility um which I think is like a beautiful vision and hopefully one that we're

hitting to but it will have some hiccups in between you know on a person on a personal perspective um I know one one thing that that I'm actually quite excited about is when when a when agents and humans unlock so much productivity because you know you're boo basically a human is boosted by by these Agents from a productivity angle and maybe by some information angle as well the acceleration curve of innovation will just progress tremendously right and and then the next question is actually then what's what's rate limiting a civilization's growth if productivity is at maximum and and knowledge sharing and communication is at maximum right and then we realize that okay it's actually just pure resources right population will grow and then it's just

pure resources right so actually the next Frontier that we like likely unlock right so like today we are like what indust the Industrial Revolution 7.0 and that's that's AI right the next one I think that's going to happen it's actually cybernetics where the only problem is the latency between a human communicating with an AI because I need to still write my thoughts out and that latency will start to matter um when itics are in place you remove that latency and suddenly again productivity exponentiate um communication and information loss uh re loss reduces communication exponential and then the the next Frontier will actually be just space exploration man we will be like we' be like Captain spook uh we on our Enterprise going to New Frontiers saying

hi to aliens right and it's likly good to be the case because you be finding new resources out there to cope with with the exponential growth of a human population and I think I I really wanted that happen right because I think that's going to be the next the next thing I work work on it's like how do we explore space well well I'm uh I'm personally very excited that I have my agent fund sponsored by you up and running so I can pay for my cybernetic subscription and my neuralink and all that so appreciate your sponsorship Jensen maybe one last question to close out but thinking through the impact of AI and like how humans interplay with that the inflation the deflation the unemployment the space travel the creativity like I don't know where where do you think like you'll be

spending your time in in 5 years like if we hit AGI with Sam Alman this year like that he's mentioned like what would you do like would you have enough money to do space travel like would you I don't know like where would you spend your time like yeah no I I think it'll be okay so so so so the honest Next Step um will actually to be going will'll be going back to my root so I have roots in in biotechnology this is what I studied back at Imperial right and and and that Vision back then was actually around how do you connect a human biology to to a to a computer it's basically what neing has been doing and and and this inspiration came actually even before

that right um sidebox and stuff like that so I think that likely would be the next next thing because I think for me what what motivates uh uh me is like always thinking about what is the Nea of civilization that you want to be working on that can that can create this kind of exponential growth right so we are in in we are in a state of of AI agent because we know that that for a fact is going to be that liver for an exponential growth curve in productivity and the next if the next bit is cybernetics right I think that's where there a field that we likely want to tap into and I think it plays very well into into my first look yeah I'm going to I'm going to Parlay this back to our space because I'm I'm curious

your take but there there are people that say that um in that in that world and in that future um we'll never be able to trust you know an open AI to to rule our world and Rule our lives especially if we're connected something to like a neural link right like you want this Tech open source you want audible you want things like that y do you think that at the end of this podcast post conversation do you think the Arc of technology and Humanity favors open source and crypto AI or favors the centralized Labs I think I think it would lean towards a decentralized environment because of trust yeah yeah no I this as simple as that it's a good answer Jansen thank you so much for for coming on the podcast and pondering durian for for co-hosting

um as a disclosure I own virtuals uh tokens so just want to make sure that's transparent and Jensen thank you so much for coming on pleases M sir always good to jam yeah great to meet you man