![[AUDIO] Yash Agarwal: From Hackathon to Ecosystem--How Send AI Is Fueling Solana’s AI Ambitions Artwork](https://www.buzzsprout.com/rails/active_storage/representations/redirect/eyJfcmFpbHMiOnsibWVzc2FnZSI6IkJBaHBCSHhUNXdnPSIsImV4cCI6bnVsbCwicHVyIjoiYmxvYl9pZCJ9fQ==--faddd9c24e07ae20e2d88c277cd5857571671f99/eyJfcmFpbHMiOnsibWVzc2FnZSI6IkJBaDdDVG9MWm05eWJXRjBPZ2hxY0djNkUzSmxjMmw2WlY5MGIxOW1hV3hzV3docEFsZ0NhUUpZQW5zR09nbGpjbTl3T2d0alpXNTBjbVU2Q25OaGRtVnlld1k2REhGMVlXeHBkSGxwUVRvUVkyOXNiM1Z5YzNCaFkyVkpJZ2x6Y21kaUJqb0dSVlE9IiwiZXhwIjpudWxsLCJwdXIiOiJ2YXJpYXRpb24ifX0=--1924d851274c06c8fa0acdfeffb43489fc4a7fcc/yash_agarwal_square.jpg)
Block by Block: A Show on Web3 Growth Marketing
Each week, I sit down with the innovators and builders shaping the future of crypto and web3.
Growth isn’t a sprint; it’s a process—built gradually, step by step, block by block.
Let’s build something incredible, together. All onchain.
Block by Block: A Show on Web3 Growth Marketing
[AUDIO] Yash Agarwal: From Hackathon to Ecosystem--How Send AI Is Fueling Solana’s AI Ambitions
Episode is brought to you by Infinex. Experience Crypto designed for humans.
Get your free account today: https://app.infinex.xyz/?r=B2KSQJ77
-------------------------
Summary
In this conversation, Yash Agarwal, co-founder of Send AI, discusses his journey into the Solana ecosystem, the motivation behind launching the Solana AI Hackathon, and the subsequent development of Send AI and the Solana Agent Kit. He emphasizes the importance of building a robust AI ecosystem on Solana and shares insights into the future of AI agents and the Send ecosystem, including the role of the Send coin and various upcoming projects.
Takeaways
Yash Agarwal entered the Solana ecosystem in 2021 during a bull market.
The Super Team community played a crucial role in his journey.
Solana's active developer community attracted him to the ecosystem.
The Solana AI Hackathon aimed to position Solana as a leader in AI agents.
Send AI was born out of the need for a focused AI initiative in Solana.
The Solana Agent Kit has seen significant adoption among developers.
Send AI aims to grow the Solana AI landscape exponentially.
The Send ecosystem includes various products, including games and tools for developers.
AI agents are still in their infancy but have great potential for future development.
The Send coin serves as an attention token for the ecosystem.
Timeline
00:00 Introduction to Yash Agarwal and Solana Ecosystem
02:53 The Rise of Send AI and the AI Hackathon
05:55 Building the Solana AI Ecosystem
09:01 The Launch of Send AI and Its Vision
11:49 The Solana Agent Kit: Features and Adoption
14:55 Expanding the Send Ecosystem: Games and Apps
17:51 Future of AI Agents and Their Market Fit
Follow me @shmula on X for upcoming episodes and to get in touch with me.
See other Episodes Here. And thank you to all our crypto and blockchain guests.
Okay, we're rolling. Yosh, co-founder of Send AI. Welcome. Yeah, thanks for having me excited to be on this podcast. I am too. I reached out to you and the team a few weeks ago. I am trying to get more projects from the Solana ecosystem on the pod and so that the public can learn more about what's going on. And very excited to learn about what you and the SendAI team are doing. Before we get into that, we'd love to hear about your background. How'd you get into crypto and specifically, of all the ecosystems you could become part of, why did you choose Solana as your home? Yeah, so I entered Solana Eco-istorm back in 2021. It was like at the peak of bull market in 2021. So that time I was looking around different Eco-istorms and Solana had a community called Super Team. So it was like a very new community and it started with Super Team in India. So I was one of the first 20 members, early members there and I started by doing some bounties. So in bounties you have to write a deep dive essay on a particular topic. So that's how like I got started in the Super Team ecosystem or the Solana ecosystem and since Super Team was exclusively to the Solana so obviously via Super Team like getting access to anyone in the Solana ecosystem was much easier because Super Team was heavily focused on educating about Solana to all the developers and also all the writers as well so I think yeah Super Team was one of the main reasons I got into Solana and in general I think that time it was that like FTX was one of the like major supporters for Solana And Solana had like a lot of good projects and I think I actually remember Serum, Mango and all of that was a big project on Solana. that's what like attracted me to Solana that there was like all the actions happening and so on. So yeah, that's the reason like I started with the Solana ecosystem. Yeah, I remember during that time, at the time Solana was doing a lot of hackathons. And one of my friends was actually the community manager and he, Austin, he was the one who started all the kind of got started creating all these hackathons around the world for Solana. And I was amazed at how, how active all the developers were. And I had been very active in the Ethereum ecosystem with DevCon and all the ETH global events for Ethereum. And I hadn't seen as much excitement outside of Ethereum until Solana and I was so surprised. Let's, I'm curious, so Polygon has a very active developer community and a lot of activity in India. Outside of Polygon and Solana, are you seeing other ecosystems that are actively trying to recruit developers in India? Right now I'm not very much updated with what's happening in India, but in general I see like other chains as well avalanche aptos and all like trying hard to get again Indian developer mindset. I think is recently monad also started like building a developer community. But I think like right now I would say Solana is one of the like top chain of choice for Indian developers and in general Indian developers are generally someone who likes to like work for different companies. and there very few projects like which are coming out of India as such. So yeah, I would say like people just like see like wherever they are seeing like more projects about openings, they generally like to go into that equation. So I would say it's less about community and more about job openings when it comes to developer interest in a particular chain. And yeah, like there a of different, different equations which are doing pretty well in India right now. Yeah, let's transition to Send AI. From what I understand, Send AI started with the first Solana AI hackathon that you and I think a couple of team members kind of were friends kind of got together. And I want to read something that you had written back in October and then maybe use that as a starting off point for a conversation. And so the title of this post is the Solana AI Hackathon story, Blood, Sweat and Tears. And then you talk about how Coinbase, the Coinbase development, what is it? The Coinbase development platform. then you end one of the intro with this statement. You said, while the base team was in full throttle mode for AI agents narrative, in contrast to Solana, No one in the builders community was talking about AI agents. The Solana foundation was sleeping on it and it's understandable given it's slow, less than 30 folks and archaic. AI agents are 2025 biggest tech narrative and seeing Solana lose the agent race deeply bothered me. I had a call with the Solana foundation, but they were still largely in denial and unconvinced about the AI agent meta. And I did not want Solana to lose. Wow. So tell us about that because that is talk about mission driven. You're not just in it to make money. You're in it to make Solana win and to contribute to the space. Like what's driving this? Yeah, so as like Solana was one of the first equitums I entered in crypto so from 2021 to 2024 I would say like I grew up in the Solana Equitum like I made a lot of friends here and that's where like I also learned a lot of things so obviously like it was like by by 2024 I had this attitude that I have to also give back to the Solana Equitum and at the same time it also gives me a chance to like be one of the leaders in the Solana Equitum So, and just before the slana AI hackathon, we also conducted a event called slana consumer day. And I've been a very big proponent of like just making sure there are more apps in the slana ecosystem and there's more focus in the consumer space and so on. But one thing I saw that slana was lacking was that it wasn't like a leader in the AI agent space. like someone like base was very much focusing on AI agents and Coinbase particularly was like very much a big proponent in the AI agents and they're still one of the biggest proponents of AI agents. So I was like, if we don't end up like being one of the early movers in AI agents space, then there's a high chance that base becomes the leading chain and Solana doesn't end up becoming the leading chain. And Solana had luckily all the capital velocity because of all the meme coin mania and all. So all the top AI coins at that time like Goat, like Goat was a top meme coin. So that was in Solana. So was like Solana has that potential and obviously as a tech it's very good and all of that. So like if you combine everything, it's a one of the best change suited for AI agents. So like what can I do to make sure that Solana is one of the number one chain for AI agents. And yeah, like that's how like I came up with this idea that maybe like a hackathon focused just on AI would be really great because before that there has been a of Solana hackathon like there's a Solana official hackathon. every like two times in a year. So they but these are not like focus on a particular niche. So I was like there's a need for a niche hackathon, which just focus on AI agents. So like they are a lot of developers and a lot of teams start thinking about AI agents. And I didn't have even expectations that there would be like lot of good projects coming out of hackathon. But I was like this also good time to like keep talking keep builders talking about AI agents and they will start thinking about how to build AI agents and so on in this space so yeah that was the motivation which came here and another thing was like earlier lot of people were in denial mode so that basically fueled the fire that now I have to do it and I'll have to prove everyone that like we need a AI hackathon and more than AI hackathon was more about AI agents narrative hackathon was just a medium to convey the narrative that I want to enforce in the slana ecosystem but it was more on this just I wanted to do this just to prove the point. So yeah, early first it started with a rational point, but later in it become like just a just I wanted to do this to prove the point. There was no other motivation because rationally didn't make sense to just go against the odds and do a hackathon. So that's how it started. Well, tell us the story after that. So this drove a fire in your belly to start this AI hackathon and start convincing the Solana ecosystem and the Solana community, hey, we need to pay attention to AI agents. We can't let base beat us. We need to be part of the story. And so you had the hackathon and then tell us what came out of it. yeah, so after I decided, okay, we have to do the hackathon. I talked to Solana foundation. They were not completely in denial mode, but they were, they weren't of not very high opinion of it that like, okay, like this is the need of a thing or even if, if AI agents are real or not. So I was like, okay, like let, let me just post it on Twitter and then see like how the community reacts to it. So what I did was like, I just made a post on it. So I like took a help from one of our design agencies. So this design agency designed entire branding for me for free just because they wanted to work with us because previously they have worked with us and they were like we just want to be part of it because they were of the opinion that this would become big so this is a very good opportunity for them to also sign up in the equation so they designed me a poster here and then I just posted on Twitter in the Twitter post it was that announcing Sulana AI hackathon And it was written $100k in prizes focus on AI agent space. And I just wrote one like reach out for partners and sponsors. And when I posted it out, initially, I didn't get like a lot of traction for the first one out, but then it started blowing up like a lot of people organically tweeted and obviously Solana Handel also, or co tweeted it and I knew Solana Handel so kind of I took help from them as well that like, can you just like co tweeted because I'm doing this for the community. and apart from that I didn't have any partners or anything so even 100k dollars in prizes that was like more of a bluff because I didn't have any prize money obviously I had this conviction that I would be able to raise it but I didn't have any prize money myself and I was like okay if something goes bad then I can just like put some money from my pocket or I would just arrange something but I had this conviction that people will support me so just tweeted it out and when it went out like it went to thousand likes and Obviously it got a lot of traction. People started talking about it. And luckily like over the next five, six days, I started like receiving a lot of DMS that I, want to sponsor this hackathon. And then the next five, six days was all about like just making sure what are the different tracks, which would be there and how can we fit it? So I just made a one page memo or just like we make for like, let's say investment memo. So I made a one page memo memo for hackathon and just send it to everyone who's whoever was DMing me. And yeah, in the next five, six days, we aimed of raising $100k but we ended up raising$280k and we ended up raising from Jupiter, Gito, Helios, like all the top teams in Solana Equestrum and then also AI16Z, Zeraburo, so all the top AI teams as well and plus VCs like Multicon Capital and so on. So at the end of the day, like everyone came here to support me and another interesting part was that I wanted to collect all the prize money in advance. So it had like two particular reasons. So one was that obviously that once there is a prize money and I don't it becomes very easy to disperse funds to hackathon participant as soon as they win. And another one was whenever someone gives prize money in advance, they have a skin in the game. So once they've given the prize money, they would be obliged to hear amplify as much as they can. So I just like took everyone's money like in advance and I was like, if you don't pay by tomorrow, like we won't have you as a sponsor. So and like I went through all the finance teams and I think everyone paid in like just five six days. So everything was very quick and yeah by by just one week we already had the prize money ready and also like all the sponsors ready all the tracks ready and then it was just like going and announcing the hackathon as the registration starts. So yeah, that's that's how like I actually ended up doing this and Surana foundation at the end was just one of the sponsors and not even the biggest sponsor. was the fourth, third or fourth biggest sponsor. And we, in total we raised from more than 30 plus sponsor and the median sponsor amount was just $5k. So which shows that it was very diverse set of sponsors and not very few sponsors putting a lot of money. yeah. And obviously like we kind of disbursed most of the prize money to the participants itself because yeah. And then like we just wanted to make sure that participants get a lot of it. So yeah. I think that's amazing. think a lot of the, you know, when you're introducing kind of a getting an organization like the Solana ecosystem, the community to pay attention to something different and new, you know, changing hearts and minds is always hard, you know, and a lot of crypto we, you know, we, over index, I think on the technical part, but really the hard, the hard things are really kind of the soft, the soft qualitative aspects of changing hearts, changing minds. convincing people you should pay attention to AI. This is not a fly by night thing. This is going to be around and we need to take it seriously and Solana needs to be a leader in the space. And I think that, you you, it sounds like you were able to get the organization, everybody to kind of pivot and think, Hey, we need to be, this, we need to take this seriously and to get all these organizations to pitch in and be part of the hackathon is, you know, they have skin in the game. That's so smart. You did really great. Now let's pivot to SendAI. That was kind of the beginning of the story. And tell us about how you founded SendAI and tell us about the, I guess, the target audience for SendAI and its kind of key features in the development kit. Yeah, so when we started Solana hackathon, we didn't have Sendai as the focus or anything. Obviously at the back of the mind we had that we want an organization, but initially it was me basically doing all the comms for the hackathon. But then I realized that like we need an organization which can be like kind of the anchor of the hackathon or the host of the hackathon. So that time we started Sendai. So we already had a coin called Sendcoin. So we started sending on, think six, seven months ago. And that was like started as more of a communal token as a meme coin, but later on, like it started like being more of a builder coin. So we already had a gaming arm called send our kid. So we thought like, let's also create an AI arm called send AI. So this end, I would basically initially be the host of the hackathon and then just like be like a development lab, which is like building a lot of, uh, AI agents and different, different, uh, AI products in the slana ecosystem. the vision of Sendai has been very simple that we want to accelerate the slana AI ecosystem. So we feel that if you grow the slana AI landscape by 10x or 100x at the same time, we will start accruing value because we are here growing this pie and our entire focus has always been to grow the slana AI because if you feel that like right now the space is very small, obviously the tokens and all are still a bit bloated, but the space in itself, the total addressable market is very small. So we want to make sure that we grow the space and in the same time we keep accruing all the value. So that was the vision like Sunday. I started with and obviously Sana AI hackathon was initially the focus. So during the Solana AI hackathon, we also wanted to make sure that developers find it very easy to build AI agents particularly on Solana. So we started building Solana agent kit which was essentially a toolkit for connecting any agents to Solana. So in that like initially we just integrated or two three protocols, but before releasing that like I've just been to all the Equestrum teams and I said that we are thinking of releasing this would you like to contribute? So I went to radium I went to archive I went to drift and all the big protocols like Metaplex and so on and I told them why don't you contribute to this kit and everything what everyone was very enthusiastic because we have already like organizing slana AI account and we were pretty much in the news and Apart from that like everyone already knew us in the slana equestrum because we have been here for quite long time so they started contributing to the kit and When we when we were about to release this like we already had 12 13 protocols integrated and everyone was like community contributed. They started contributing themselves. So that way when we announced the Solana agent kit these people also since they can contribute they had also skin in the game. So they were they also amplified hard and when we released Solana agent kit, it was a huge success. For example, the post got more than 1 million views more than 2.5k likes and everyone was talking about it because This was like one of the first toolkit on a blockchain level, which is like just focus on a particular blockchain. So earlier like Swinebase had this agent kit, but it was exclusively on base, but it had just three, four protocols integrated. But here we were releasing with almost 13, 14 protocols integrated. So that's how like Solana agent kit was released. And currently we have around 30 protocols integrated and we have more than 90,000 downloads around 1100 stars on GitHub and 600 folks. So from the developers and adoption point of view, it has been growing very fast and a lot of builders are building in the building using the Solana agent kit. Even in hackathon, we had around 60 plus projects which built using Solana agent kit despite us releasing just three, four days before the hackathon. So that showed that like people wanted to build things and here like we don't have any competitor or we don't have any replacement because if you want to integrate any Solana SDK is very easily Solana agent kit makes it very easy to connect your AI agents. So yeah, like we are kind of exclusive to Solana. We don't have any competitors or any alternative as such. But at the same time, it's completely free public open source goods. So we can't directly monetize from the Solana agent kit as well. yeah, right now for Sendai, one of the main focus is Solana agent kit, but we are planning like different other products and initiatives like more initiatives like maybe like an AI accelerator. And obviously we have different regional activations which are going on. For example, we'll be conducting a demo day in UAE next month. apart from that, like, just like yesterday, we conducted a workshop with Super Team Poland. So keep conducting like different, different regional workshops and everything because yeah, for us, it's like just growing the slana AI space and keep building different products because if we have the good developer mindset and the community mindset, everyone would be like eager to use our products and also our toolkits and whatever like we keep building. So yeah, that's been the broadly the vision of Sanjaya. So the Solana agent kit, you're not monetizing it. It's a free public good that you're giving to the Solana ecosystem. That's amazing that you guys are doing that. What apps have you seen using the agent kit that you're excited about? Oh yeah, so one of them can be would be new. So new is like kind of a chat GPT for Solana where you give instructions and then it does certain things. So in fact, like we saw that like a lot of people were building kind of you can say chat GPT rappers or a chat interface where people can do Solana accents. So we also made our own such type of thing where we open sourced it completely. And now you can just like install it using a packet. So for example, you just like do npm install create solana agent and it will install that package and in just five minutes in less than five minutes actually you can build your own LLM wrapper in that like you can say that LLM to create a orca pool or like trade on jupiter obviously even like long or short on drift so any actions which are possible on solana agent you can do that obviously it's not very production friendly that like it won't work 100 % of times because there's still times when this model is hallucinate And since these models aren't very optimized for pool calling yet. So this is one of the problems that we have identified. But I would say at least for the 95 % of the cases, it will work well. So this is one of the package that we built. So like anyone can build such types of things and obviously keep improving. there's another project called Hive, which is like kind of a agent orchestration. They have been using Solana AgentKit. Apart from that, like they're lot of other projects as well which have integrated Solana AgentKit so people are also using Solana AgentKit in some partial forms so for example they are integrating few SDKs from Solana AgentKit not the entire kit because if you use the entire kit then there's a chance that tools might start hallucinating because there's a lot of SDKs and LLM might get confused so yeah I would say like agent orchestration is like one of the good use cases that I've seen so far LLM wrappers are obviously a good use cases as well And then like there are like different different DFAI agents as well. For example, yield optimizer agents. And then there's like LP agent, which is using Solana agent kit to LP on Meteora. So essentially if you are building any agents on top of these protocols, you can use Solana agent kit because right now we are the only tooling layer for Solana, which covers like a wide variety of protocols. And we also have AI integrations like LAN chain, etc. So that like this frameworks already have the tool context. So you don't have to worry about the giving context and everything. So you just need to decide which LLM model to use. And till now we have figured out that cloud has worked the best because it has the best tool calling among different AI models. So if you're using cloud and line chain and then this different tool tools, like you can build a very good AI agent or LLM or even a chat interface, which can work for most of the cases. So this has been our observation so far and what people have been building. on top of it. I met with, I don't know if you've heard of Open Ledger. I met with them, interviewed them last week and you know, their approach to kind of decentralize AI is very interesting. And it's, for the most part, I think it's chain agnostic. And so I'm curious since Send AI, the Solana agent kit is dedicated, but can it also be used by applications and other chains? not directly so we have some bridge integration for example we have my and finance integrated and then we also have deep bridge integrated and soon we'll have a much deeper integration with vermouth as well where like you can do intense sort of transactions like they just release the intense protocol so we like we are integrating that deeply as well so obviously like we have such cross-chain integrations so you let's say when you agents is using any is doing any transaction on base and they want to send it to slana as well that can be done And if any agents want to let's say fund wallet from base or polygon or different other change they can do that as well via Solana agent kit. But currently we are like strategically focused only on Solana because sender started as like growing the Solana ecosystem. So obviously we might expand to other chains and maybe like add different capabilities, but we want to be vertically very focused on Solana because we feel that like in Solana there's a lot of improvements that can be done. And if you focus just on Solana and we believe that like at then Solana would be the execution layer for agents. So even call all the all the agents which are built using Solana agent kit as the execution agent because these are executing on chain unlike other AI agents which are like let's say tweeting on on Twitter or basically just like doing doing live streams and so on. So we want more and on chain agents which can execute on chain and we want Solana would to be the executioner layer. So that's why like we also named it Solana agent kit and We just want to make sure that like we in short any agent which is coming whether be a web 2 agent or anyone who is like not even in crypto, but this would be able to connect to Solana and connect all the protocols and this Solana agent is wallet agnostic. So you can connect any types of wallet like embedded wallet or even like your phantom wallet and so on. So at the end of the day Solana agent could be completely wallet agnostic. It's just like it needs to be a Solana wallet or in fact even it can be even our SVM wallet for example, there's a layer two on Solana called Sonic SVM. So we essentially cloned Solana agent kit to that. So now it's a Sonic agent kit as well. And we didn't have to do any any changes like it works perfectly fine on that as well. just that those protocols need to support that. So on the wallet level like all the transactions signing everything works perfectly fine on any SVM. So it's not that we are just focused on Solana assets, but it's like we we can be operational in any SVM environment. Yeah. So we've been talking about the Solana agent kit and as kind of one part of Send AI. But the Send ecosystem has lots of other products and offerings in it. Send.fun is one. Can you tell us a little bit about Send.fun? yeah so send actually structured in a way that it has a token called send token so which again like has been operational from the last seven eight months and like we had seen a lot of ups and downs in your token and obviously like it's like kind of you can say it's a way to speculate on the send ecosystem in general and then there's a gaming arm called send arcade so send arcade is like heavily focused on building different fun games so for example they have a game called sneaks and ladders where you roll a dice and you just like keep playing so if you fall in a ladder then you get like some meme coins if you fall in some snakes then you basically climb down the ladder and so on so essentially it's like kind of a fun game like but it has like a bit of speculative component because you are getting meme coins and every time you roll the dice you have to spend some soul and in most of their games how it works is like you play via soul but you end up winning send so every time a user plays a game the soul is converted into send and if they win like they get send as a winning so that way every game has an inherent by pressure to send so that that's how like send also accrues utility here because it has an inherent by pressure and there's a high chance that anyone who wins send doesn't directly sell it so we can say that like it's a good trojan horse for onboarding people to send equestria mess with and till now like this games have done more than 170,000 transactions and around $1.5 million in volumes and we didn't like spend anything marketing. Most of our reach has been organic and it has been coming from channels like backpack because we are natively integrated in backpack as blinks. And then we also like deployed on eclipse, which is another layer to an Ethereum. And then apart from that, like we also have like other distribution channels from where people keep getting the word of mouth is also a decently strong here. But yeah, like obviously this is like a very good first step and kind of a good cash cow as well because it keeps generating small revenues here but the game here would be the bigger like for example, we are now building a no code transaction builder. So example, if you build any game like let's say pinball or any sorts of normal games, you can just integrate this transaction builders and now it can be developed converted into a web 3 game because now this can have all the soft functionality or if you want to integrate Metaplex NFTs that can be also done. And if you just want to have, let's say, pay per transactions, you can do that and everything would be no code. So basically like you just bring any one of the games and then like attach all the transaction logic and you just deploy that. So this is like one of the projects that we are building. Then we also thinking of a unity framework for Solana, for example, so that like all the older game developers, are building using unity, can integrate Solana easily and do some Solana actions. And we also building our own app called Arcade app. because one insight we got was right now we are relying on different distribution channels because all these games are essentially blinks inside different apps. So it's like kind of mini games. So we thought that it's better to also have our own app where you have options to do lot of mini games. So right now we have six mini games, but we want to expand it to 10 mini games and and in future even hundred mini games and we want to also enable developers to keep building this mini games and there will be a single app where people can browse all the mini games and there's a single wallet. and they can use this wallet across different games. So that way we kind of own the users and at the same time, all this user will have a username and we also start building a social graph here. So it would be similar to what Facebook was in 2013, 14, well, there are a of different mini games and at same time it was a social network. So it's like kind of inspired from that, but completely built on Solana and using Solana as a rails for the entire payment things and also on the speculative layer. So maybe like most of these games can even have bonding curve and so on. So it will be fun speculative apps like targeted for users who just want to have fun and do small ticket size transaction. apart from that, like on the send request, we also soon building a Solana app kit. So just like we built Solana agent kit, one of the insights that we got by building like a lot of apps ourselves and a lot of consumer applications has been that. We can also enable building consumer apps on Solana much easier just like agents. So this would be a app kit which will be focusing on building apps very easily. So it would have reference implementation of bonding curves, then NFTs and so on. their developers can just plug and play and as like AI is becoming much more advanced. I personally feel that the cost of producing apps would be significantly lower. So as the cost of producing apps are significantly lower, people would start building this apps in just two or three days. And here we are providing all the plug and play architecture and all the Solana integrations and it will also enable any people whether be from a traditional background or anyone who doesn't even know about Solana. There's this like plug and play everything and build a Solana consumer apps very fast. So that's also one of the vision that we are building for Solana app kit, which again, like initially would be complete open source, but we also want to monetize. later on, like we start building our own platform and that ties to both apps as well as agents that even on the AI agent side, we would also start building our own platform which starts accruing value, but we still feel that like agents are not doesn't have PMF or doesn't have a good utility so far. So we are just waiting for that because a platform with agents like which doesn't have any utility doesn't make sense. So for now, like we are just like focusing on building the infrastructure and when we feel that now is a good time to start monetizing and build our own platform. we will start doing that because we at the of the day have the entire ecosystem ready. But in general for send ecosystem, the focus is entirely on apps and agents. And another interesting thing which we are doing is since we are talking to a lot of builders day in and day out and we have a of a lot of different initiatives like AI hackathons and then other sorts of events and so on. For example, we also host of Solana consumer day. So we thought that it's a very good way to also monetize all the all the cool projects that we work with. So we'll be also launching send ventures, which would be initially like an angel syndicate where different Solana founders can come and invest in the early stage startups. And obviously it would be at a very early stage where we invest like 200k dollars for let's say 5 % or 10 % and we help them go through the entire journey and like just like keep helping us as a builders as well. So that's another initiative that we are planning. Obviously, long term plan here would be that like we raise from we raise a proper fund from LPs and so on. But right now it would be starting as an angel syndicate. So yeah, we are kind of like just trying to build a lot of different projects inside the send ecosystem and everything is tied to a single token, the send coin. So that's how we have structured the entire send ecosystem. Obviously, it's a very ambitious play and we are still an experiment. It might work, it might not work. But we love doing experiments and our entire team is like focus on doing a lot of experiments. So yeah, like I just always say that send is an experiment and send is mostly our tension token that as we keep building more products and do keep doing different initiatives, there would be more attention and send will keep according value of the send will be integrated wherever it's required and there would be like revenue buybacks and everything and so on. But I never called it extend is a governance token or a utility token. It's simply our tension token because we have a lot of products and initiatives and everything is tied to one token, which is the same point. So yeah, that's how the entire send ecosystem is structured. That sounds great. mean, you have a lot going on and it sounds like you're actively growing the SEND ecosystem and it's represented by this attention token, the SEND coin. Let me ask you one more question around AI agents in general. I think AIXBT is probably the most well-known agent, but it doesn't really have product market fit. It doesn't really do anything other than just replies. Where do you see kind of like the next steps in terms of like AI agents and how can they become more useful and what I guess what's your vision? Sure, so I think there's a reason that AI agents currently doesn't have like any PMF or anything of sorts. think like AI XBT is one of the best agents out there from a summary point of view. example I would say AI XBT is a very good CT summary agent or even you can say a bot so that like it summarizes what's happening in CT and then like correlates that with the prizes and then gives a very good replies and all of that. So I think like at the end of the day it has gotten attention. So I would give full credit is that eixbd is like one of the best agents out there and something which has partial PMF because it at least it's getting what it's meant for that is getting the attention but I think like it's not purely a agent because it's not executing anything if it was executing trades after what it has been analyzing then I would have seen said that it's a proper agent because it also executing on chain. So what I say crypto cross AI agents or on chain agents are anything which executes on chain. So an agent agent is simply basically an LLM plus a bot. So LLMs essentially allow people like any agent to think and then bot is the executional part. So I think like bots currently in crypto has a very good PMF. you can say like bots in crypto does more than five million dollars in revenue each day like bong bot, trojan bot and then like lot of different sniper bots which people deploy. I think bot has a very good PMF in crypto. It's now about that we make this bots much more intelligent and smarter. So that will be like allowing this bot also to think autonomously. So I feel that soon like we will start getting like some trading agents, which is like collecting alpha from different sources and then executing engine. They're still not there yet. And I think like one thing like a lot of people have been here ignoring the fact that we don't need like a lot of AI agent frameworks. AI agent frameworks are already there in web2 like Langchain is a very good framework like For example, we have seen that like most of like Lantern is one of the best frameworks to be used with Solana agent kit and not any crypto frameworks. The reason being that like Lantern has is very much focused on building the agent workflows. So I feel that like people in crypto should also start thinking from that point of view. How do we build an agent workflows? For example, if some agent does yield farming on one protocol and then after that it starts in deposit and stakes that yield in some protocols. So it's like basically a chain of a actions when anything is done on chain. So I think like and then everything should be taught of an agentive workflow and this existing web two frameworks work perfectly fine. I think people here are missing the point on how do we get the quality data because one thing is that model all the AI model aren't like very much like trained for blockchain sort of data like any any model that cloud or even chat gpd etc. did doesn't understand like how to read the blockchain transactions or how exactly what's exactly happening in a blockchain transaction. So I feel that right now a lot of work has to be done on the model layer as well. For example, fine tuning a model just on blockchain qualitative as well as quantitative data for quantitative. I mean transactions data that the model should be able to understand what's what's in a transaction and how exactly agents do a transaction. So I think like once that context is there and models are very much optimized for tool calling. then we would start seeing like lot of good agents which are able to think autonomously read the blockchain and then like take actions based on that and and that's the reason that we are currently very much focused on the tooling layer because we believe that that's the real mode of crypto cross AI that we have access to like AI agent has access to wallet and once it has access to wallet it can access all the protocols and different sort of keep doing different sort of agentic actions for example we also want to ensure that our agent could has all the actions required for making purchases or even on Amazon. So we will be soon doing a integration which will allow AI agents to pay for its own subscriptions and all of that because it would have a virtual debit card. So we are talking to a company where like we want to make this a possibility where this AI agents cannot not just execute on Solana but also like keep doing real world transactions so that like we make it like not just on chain agents but also like agents which can do transaction in the real world. And one more thing that we felt that our agent can't be autonomous here because he has to pay for its server bills and most of the server bills happens via card. that's why like we are doing this integration. So for us, it would be like just like keep building the tooling layer because once you build that, once you believe that there's a good tooling layer, a lot of good on-chain or executional agents would be possible. And that's the vision that we are building for. And I'm also excited about that. We already have bots in crypto and that has PMF. It's now time to turn bots into like autonomous bots or you can just say agents where they not just like do actions based on a certain set of parameters based on certain sets of programmatic actions but it also is like starts thinking and start doing this it has its own funds and start doing like all of this but I feel that agents are still very early but by the next year or so like they will start doing like significant amount of blockchain activity and I think like we already have some trends there by via bots where bots currently do lot of transactions on solana so it would be like just replace via agents which are a smarter version of bots Yeah, I think I heard it once where any serious technology starts out as a toy. And then when you take it seriously, it goes by really fast. Eventually it becomes, you know, it takes over and becomes ubiquitous. And I think probably the ancient space is like that, where right now it's, you know, we've had bots for a while, but now we need to make bots intelligent using AI. And, you know, you, the SEND team, you guys are creating these really good public goods that now developers can take advantage of and hopefully we'll be able to see some really great consumer apps taking advantage and with that being powered by Send Solana agent kit. Amazing. Well, Yosh, thank you so much for taking the time to speak with us. I will share everything, all of the URLs and links in the show notes. because I want people to know about this. It would be great if more people use the Salon Agent Kit and got involved in the Sun ecosystem. Yeah, that would be awesome. And yeah, in general, like we keep building different products for developers like Solana AgentKit and then like Soon Solana AppKit and we're building a lot of things because here like we have a simple vision that we want to accelerate apps and agents on Solana. So we just like making sure like we are building more apps ourselves, agents ourselves and then also enabling developers to build more and more apps and agents. So we have a simple vision that we want to accelerate apps and agents on Solana. That's the vision of CENT currently. That's a great vision. I support that. Thank you so much, Yash. Awesome. Yeah. Thanks for having me. was fun chatting. Yeah.
