Saltorious.eth · Among Traitors

[0:07]So welcome, welcome everyone. I hope you guys are doing amazing. Um, I'm here with another better call Zaal Yaps with saltorious. Is that how to pronounce it? Yeah, saltorious. You got it. saltorious. How you doing today saltorious? I'm good man. Uh, thank you for having me on, dude.

[0:23]Very cool. Yeah, no, it's, uh, awesome to have you on, um, been, uh, trolling the farcaster streets and, and saw you a built something that seems a really interesting, that I wanna learn more about. But then b um, you know, you're a builder here in the farcaster streets trying to get. The word out that, you know, farcaster not dead, and there's all these awesome builders doing cool things that kind of can right now only be done on farcaster, which is pretty cool.

[0:51]Oh yeah. So, yeah, I'd love for you to give an intro. Um, tell us a little bit more about yourself and, uh, and yeah, we can get started. Yeah, for sure. Um, so go by saltorious. Um, I've been in crypto for. A good one. Now, since 2017, um, I started out kind of like anybody else, a trader speculator, Googling, how can I make money?

[1:15]Person that pops up. Um, uh, and at the time I was, uh. I was in school, school. I was finishing up a, a finance degree actually, and going into, working at, uh, uh, I was going, I was first working at Merrill Lynch and then I worked at Goldman Sachs for a while, and all the while I was like entrenched, trying to entrench myself in the crypto scene.

[1:37]Um, and then, uh, for, uh, around COVID time, uh, when everything kind of went crazy, uh, I kind of took a look at what I was doing. I was at cold make the time. I was, took a look at what I was doing and I was kind of like, I can't do this for the next 40 years of my life. I absolutely hated every day of going into the office, and I really wanted to take a stab at.

[2:02]Um, making a career out of crypto. So I actually, the day that lockdowns happened, I quit my job and, uh, I started working on what was the first, uh, crypto venture, uh, self, you know, self. Solo dev crypto venture. Um, and then from there, it's just been a snowball. Uh, so I've been able to part, uh, to uh, contribute to a few DAOs.

[2:28]I contribute to a DAO called One Hive for a little while. Um, what was one Hive's uh, goal? Uh, one hive was basically, um. Creating decentralized products on the nos chain. So, uh, a lot of what this was like around the sushi swap time, uh, when like DEXs were really first popping off. Um, and there was nothing on like within the gnosis chain or the gnosis ecosystem that kind of served that purpose.

[2:55]So that now existed to build, you know, um. Uh, defi Farms, like when that was really big at the time. Uh, they, they had a decentralized exchange called Honey Swap. I helped work on that. They had a lending platform called Agave. I worked on that too. Um, so very much entrenched in like the gnosis ecosystem in for the first like year or two that I started like building in crypto.

[3:20]Um. And, and then, uh, from there, I, I worked at a gaming startup, a Web3 gaming startup called Internet Game, uh, for about a year. And, um, that was cool. It was, uh, they did, uh, some really cool stuff. Like they bought a bunch of blue chip NFTs, this is maybe around 20 22, 20 23. And they had like a, a, um, what's it called?

[3:44]Uh. Squid games style, like online game, uh, where like a bunch of people would, would, uh, participate in like these, like flash games pretty much that they set up. And the winners, the winners would get, uh, like a bored ape, which was really cool at the time. Yeah, yeah. Someone would play a game for so people would go hard.

[4:04]Yeah. Yeah. They were playing a game for two weeks and walk away with like a 200,000 NFT. Oh, what the fuck? Uh, so that was cool. I was there for about a year. Um, and they pivoted to another direction. And since then I've been working at, uh, a company called Laser Technologies. Uh, some of y'all might be familiar, but Laser is a dev shop.

[4:24]Uh, probably one of the biggest dev shops in crypto at the moment, and they work with a lot of the largest, um, companies in crypto right now. Uh, so I've been there ever since. I love it. I would not change a thing about, uh, my time at Laser. Um. Absolutely incredible experience. And then on the side, obviously, uh, I'm a builder within the farcaster ecosystem, I've launched a few mini apps to varying degrees of success.

[4:49]Uh, but probably the, if, um, any of your viewers have seen my username floating around, uh, a couple of bigger ones. I've made one, uh, mini app called hypeman, and then a miniapp called People's Donuts. And then. Uh, among traders. So it's kind of like a condensed version of, uh, of my journey in crypto, but yeah.

[5:10]Yeah, just love building. I love to hear that. No, that sounds honestly like a really interesting, fun adventure. I think we all, we. Individuals as a humans realize that they don't wanna do an eight to five, and some people choose to do something about it. And I love hearing that kind of like journey of, um, popping into a startup and then finding something that like really suits you and also, uh.

[5:34]Goes well and aligns well with that like builder mentality of solo deving, different things on the side, and being able to have the bandwidth for that and kind of like have something that, uh, that combines that well. So that's really cool to hear. Um, I'd love to know like a little bit more about like your farcaster experience.

[5:51]Like how did you. Get into farcaster is the decentralized social media part, uh, part that you're really interested in is the tech part is something you're interested in, just the individuals there, like where are you at now, but like also how'd you come into the ecosystem? Uh, so to be honest, it actually started through laser.

[6:09]So when I joined laser, uh. Garrett, who you might know on the timeline. Garrett, uh, is the head of crypto at Laser. And then I also work with a town who's the guy who founded emerge. Um, and they were both super, like into farcaster and like the very first thing they told me was like, you gotta get on farcaster, dude.

[6:28]Like, that's the place that that's see the place to be at moment. And um, those are two great individuals to help onboard you as well. Oh yeah, for sure. That, that, that makes a big. Positive part of it, right? There's a lot of individual Yeah, definitely showed me the ropes and like, showed, showed me what, and then also showed me the power of like what decentralized social, which you could get out of decentralized social, um, you know, a, as a builder, I've always tried to like be active on Twitter and like do the Twitter thing, but I just hate how.

[7:01]Toxic and gameable, the timeline is over there. Uh, there's a lot of noise and not a lot of signal, and like that's kind of one of the biggest turnoffs for me to be able, like to spend a lot of time there. Um, and so yeah, I got onto farcaster through the recommendation of Garrett and a town. Um, but like within like a week.

[7:22]I was like, oh yeah, this is like, just the quality of conversations that were happening all the time. Like, this is before I even thought about building anything. Uh, just being on the timeline and seeing the quality of conversations and like the actual deliberation that was happening, uh, on farcaster there was enough for me to stay.

[7:38]And then like you layer in, oh, there's an embedded wallet, and then you layer in, oh, you can build apps on top of the whole, um, protocol. And then, and then you layer in, oh, there's like a de dedicated infrastructure to be able to like. Get people's social grasp without having to spend a thousand dollars a month on an API key.

[7:55]Like it's, it kind of all sort of emerged into place for me. And then maybe three months into being on the network, I was like, oh, okay, yeah, this is, this is where I'm gonna be at for the next little bit. Um, really enjoying my time here. Um, but yeah, I, uh, I guess so I've been, I've been like pretty active on farcaster for like the past year and a half.

[8:17]Uh, I went to Far Con in Brooklyn. Which was like just a couple months after I joined, so I didn't know anybody while I was there. But Garrett and a town, you know, told me to come by and like, I was really inspired there by like what everyone was feeling. This was a, at a time when, you know, um, I think V two frames had just dropped like a couple of months right?

[8:39]A couple months before, and everyone at Far Con was like showing off the things that they had built, all the new frames they had built, all the cool integrations that they were able to. Um, that they were able to come up with. Uh, and I just got super inspired. So literally on the train ride home, I started working on the first mini app that I ever released on farcaster, which was called Warp Ramp.

[9:02]Um, and this was before, uh, you could onboard, you could, uh, fund your, your warplet or your, your farcaster wallet with, with fiat. You can do that now directly on, on, uh, within the app, but at the time you couldn't do that. So I made. Uh, the first Minneapolis was called Warp Ramp, but just connected Coinbase on ramp to your farcaster wallet and let you on, let it lets you, uh, fund your, your farcaster wallet with um.

[9:27]With, with fiat, like directly in a mini app without having to leave, uh, the app. So like that was, that's one thing I was frustrated with. Like, if I wanted to fund my mini app, I'd have to cop in my wallet, go into Coinbase, go to wherever, put my wallet in, like you switch back and forth. I was like, why can't I just do it all within the timeline?

[9:45]So that was cool. Um. Uh, got a lot of like real, really nice buzz, like right away. I think I got a follow from DWR and I got a follow from, uh, Linda, or like a lot of the big names of farcaster at the time and I was like, oh heck yeah, this is, this is awesome. So from there I that, that's when I caught like the mini app bug and I was like, okay, it's time to, this is, this is, uh, a really awesome place to be building.

[10:10]Um, and then I, and then from there I kind of experimented with a few other mini apps. Uh, some went unreleased, some went released, but I didn't really post about it. Um, and then I, uh, uh, and then maybe like, you know, this was maybe six months ago at this point, I, uh, had an idea for a mini app called, uh, a promotional mini app that I called Hype Man.

[10:37]Uh, the idea was, um. If I wanted to promote something, uh, and I wanted, you could, there was amps, there was a couple of other min apps that existed that let you buy likes and recasts, but they only let you do it from people. That already followed you and there was no surface area for quote casts. And I think quote casts is a really high signal.

[10:56]Like if you can get a bunch of people on the timeline talking about like your thing, like that's a great way to drive engagement and buzz to like whatever you're trying do. Hundred percent. So the idea of a hype man was you could pay for quote casts, um, but instead of. The person, like actually looking at and writing the quote cast themselves.

[11:18]There's an, uh, an ai, um, engine that, uh, scan their whole profile, scan the, their top cast, learned how they write, and then it would craft the cast for them and post and uh, and then you would be able to see it, approve it, and then get paid for actually sending the cast. So it's basically you can promote a cast, whoever's promoting it for you.

[11:40]Uh, they just have to. You know, say that they wanna promote it and AI will write everything for you and then it's just check and post, which was really nice. Uh, and hypeman was like the fastest growing app that I had personally ever made. I think in two weeks there was something like six or 7,000 users.

[11:59]There was like 20,000 transactions that had happened. I was like, whoa, this is. This is like, so I, I had built another, like before farcaster or before this actually. The first, uh, like crypto product that I'd ever built was a tournament platform on top of, uh, the trading card game, the Web3 trading called card game called Gods Un-Chained.

[12:22]So Gods Onchain is kind of like, uh, yeah. Magic the Gathering or Hearthstone, but the cards are NFTs. Um, and I built a platform. Yeah, you just literally spoke my language. I got into Web3 because the Tcgs Yes. And like Magic was perfect. Application was the original one, and then Hearthstone, and then I tried Gods Unchained a little bit.

[12:41]So No, I get it. Yes. And I think, I still think it's the perfect, like one of the most perfect applications for NFTs, but uh, the success the. Shortcomings of Web3 gaming is like, yeah. For another conversation. But, uh, I, I had built, the first app app that I had ever built was called GE Stakes. Um, and that was basically, uh, it started as like a place where you could go, uh, and you one v one Stakes on Gods and Chain.

[13:08]So Gods and Chain was really very full. Oh, okay. That's awesome. Yeah. Yeah. So it's very forward thinking. The fact that one of the first things they released was an API to connect to the game. So you could, you could get like. Match results. Yeah. Could see who won, who lost what cards they used. That's cool.

[13:21]I didn't know that. Yeah. And um, basically I just built a wrapper on top of it that also let you. Bet. Yeah, bet against your friends. Um, and that like was met with some initial skepticism 'cause like, I was not really known at all, but then people started using it and people started really liking it. Um, and then it, the platform kind of, there's a bunch of people from the community who came on and, and helped me, like develop it, not only from like a code perspective, but from like a community management perspective.

[13:54]And, uh, we introduced tournaments. Um. Uh, after a little while. And so it's basically the same contract layer. I love that. Yeah. Same contract layer, but then building on top tournament. And at the time, like, uh, official gods and chain tournaments were run by this website called community gaming, and they would hire people that had no idea what gods un chain was.

[14:17]How to play the game. They didn't know anything about the game, but they were moderate in the tournament. So a lot of times, ah, a lot of times there would be cheaters, people who were playing cards that were forbidden, there'd be rulings that were just not fair. There's like very little competitive integrity for like a game that's trying to be Yeah, a combat.

[14:34]Like the Web3. Exactly. The Web3 card game, like the most popular Web3 game. And there was, so the tournament platform was kind of there to solve all those issues, but then also adds like. Uh, an integrated money or integrated, uh, like payment infrastructure on top of it. Uh, 'cause that was another issue at the time.

[14:54]You would win a tournament and it would take you a, it would take a month to get your money. And like people were like, this is Web3. Whatcha talking? It takes 10 seconds actually going on. That's crazy. Yeah. So I feel like a lot of people say they're in Web3 but are just doing web two things, right? Saying it's wrapped, wrapped in some small thing that is.

[15:14]Semi tangentially related. Yeah. 10. And then, yeah. And then you can say, if you're not using blockchain for your payment system, I feel like it's like 2.5. Exactly. Or, or maybe you are using blockchain for your payments, but the but the processes that you're using are still stuck in like 2006. Centralized usually.

[15:33]Right, exactly. In some ways, shape or form. Exactly. So, uh, that's cool. So, you know, uh, so, so got so GU stakes, you know, got, so, got after the tournaments. Feature was released, uh, you know, kind of took off. We partnered with Immutable and partnered with Gods UnChained to do what was called the summer series, and we got like $200,000 worth of like, assets and cash to like give out as prizes.

[15:59]Uh, and then from there, like a lot of, a lot of the, it became like the community. Tournament platform for Gods and Chain. Like there's a official provider like community gaming, like the big dogs, but then people kind of gravitated towards G Stakes 'cause there's a better, uh, option. Um, but, uh, but yeah, so, so from there, like bringing all of that experience to like building farcaster, uh, mini apps, um, uh, yeah, that kind of, kind of Oh, oh yes.

[16:31]Uh, so the, so Gu stakes. Uh, it maybe took it over the course of like the entire time I was developing for it, I was maybe two and a half years, uh, over the whole course of like, its lifecycle. I probably got 5,000 users all in, all in like two and a half years. Uh, and then with Hype Man, it took like two weeks to get like five Ks, you know what I mean?

[16:51]And it was. Like part of it was, you know, um, people liked the idea, but also part of it was the dis the distribution mechanism of being able to like post about something and then have a bunch of people be able to see it and use it right away. And that's where that part and be incentivized to actually come in and do, and do the reply then, and not just like, come back later.

[17:11]Which I feel like is a big part of social where like you're like, oh, like I would interact with this but I don't have the time right now, or this or that. Right. So I think that's really cool. Yeah, exactly. And, and that's when you know, and I'll, yeah, I had the, the, basically that same exact thought and came to that same exact conclusion.

[17:28]And then also like the layering of just having your entire, like the, the social graph and your social identity baked into the app, having the context available within the code. Mm-hmm. Like right there without having to do anything special or crazy. And then also having to, having like. Uh, uh, like a standard that that persists across different mini apps was, is really nice.

[17:51]Um, so yeah, that's when that part, you know, that part clicked for me. And, um, to this day I still, I always say like, farcaster is the best place to build. You can, you can get from an idea to like your first a hundred users quicker than anywhere else, in my opinion. And that's why that's, that's the most valuable thing I think.

[18:10]Uh, farcaster, uh, offers builders at the moment. Dude, 1000000%. I love that. I think that is one of the biggest benefits is that like the test, like the testing portion of trying out developing an app is a lot more fluid on a platform like farcaster because of some of those things. Yeah. So that's cool to hear.

[18:35]Um, yeah, no, I'd love to hear more about, uh, what you're doing right now among traitors. Like what kind of inspired it and then. Um, we can even pull it up over on my screen and do a little demo and you can kind of share more of, uh, of what's going on. Yeah, for sure. Um, so actually, funny enough, uh. Uh, the idea for the game came from, I stayed at the base house at ETH Denver, which happened like with Luciano?

[19:02]Yeah, with Luciano. Yeah. So Luc that's awesome. I actually went by the base house on the day before, a few days before when I was still in Boulder. I, I only swung through for the ETH Boulder event, but, uh, yeah, so I just missed you, dude. Thanks. I showed up on the Sunday, so I, I was staying there and.

[19:20]There's a based house hackathon, but then there was also an East Denver hackathon. And, uh, at the beginning of the week I was kind of like, all right, I, I know I wanna build something with agents. I know I want to do something agentic. 'cause that's what, that's what the, that's what everybody is interested in at the moment.

[19:39]And I also want to just kind of like, like learn the stack. Learn, learn. You know. How to most effectively, uh, use these things. So it started out as like just an experiment. Like how can, how can I get a bunch of agents to deliberate. On the smallest instance that I can possibly run, because I have, I've seen, you know, people set up, the people set ups online, like they're running like $10,000 rigs and have like seven GPUs, and each GPU is running its own agent instance and like it's all locally run.

[20:11]And that that's, that's all really cool. But I hadn't seen anything kind of like replicated. Online. And so, uh, what I spent the first, like half of the base house week, uh, the, the week in the base house doing was, uh, basically seeing if I could set that up. And it started out as like just having 10 individual pieces of context and running through like a very rudimentary loop.

[20:34]And then passing the, passing the first piece of context to an LLM and then taking the results. Along with the next piece of context of passing that to the next agent. Um, and that, and it, it worked kind of, but it like, it wasn't really like a conversation. It was more like just agents stacking like what the last agent had said on top of each other.

[20:56]Uh, so then, uh, I had a conversation. I stayed with, uh, the dude or bribe. I don't know if you're familiar. I love him. Yeah. Love that guy. He's a goat. Love that guy. So. Uh, we, we roomed together and kind of after a conversation with him, I was like, oh, this can be done, like in a very, in a much more intelligent way.

[21:14]So, uh, I was then experimenting with like using node workers, uh, using worker threads to like run the inference and run the context management. And that was like, way better. And then he was like. What if you just threw 'em all into an XMTP chat? I was like, what if I did? And so I like started experimenting with that.

[21:33]And um, I think like on the fourth day or the third day or something like that, uh, we were like just watching these agents deliberate. I had made hi, uh, I had made a profile for like his, the DUDE account. And it was actually so funny. Like I could only imagine it was the, um. The description of like, his character, first of all was like him to a TI don't if you've ever met him.

[21:59]It was like him to a T And then the description talks about, uh, him like being obsessed with Palo Santo and making like generative NFT collections that like represent represented people's souls. And at the base house, he brought a bunch of Palo Santo for everybody and was like giving it out. And he was in the middle of making like a generative NFT collection.

[22:22]So it was like, dude, this is exactly you. And then we watched like how his agent deliberated was interacting. Yeah, we chat, we set up. So like, it was, it was all a very funny experience. And so from there I was like, I think this would be a fun game. Like what if you, what if you locked Ted agents to the room and you gave them like a game to play, like a directive.

[22:41]Yeah. And so the very first thing that I thought of is like me and my friends, like back, back home. Uh, we go to a ski trip in Vermont every year. We play this game called Salem. Uh, and Salem is essentially, uh, a little bit more complicated version of among traders, but you have somebody, people who are witches and people who are the townspeople, and it's up to the townspeople to figure out who the witches are before they kill everybody.

[23:05]So it's like, what if you just threw a bunch of agents in a room and made them play like a simplified version of that? Um, and that's kind of where the idea was born, born from. Uh, and so the game, uh, if anybody's unfamiliar, uh, the game is you spin up an agent, your agent's based on your farcaster profile or your eventually your Twitter profile.

[23:28]Once I figure out how I'm gonna pay to use the Twitter API. But, uh, it's based on your, at the moment, it's based on your farcaster profile, and it's not. A copy of you, but it's like a personality that's based off of your online presence. Um, and, uh, once your, once your agent is birthed, you can join a lobby.

[23:50]Oh, cool. You got it up on screen here. So once this is what you see, yeah, that's what you see. You can sign in, uh, and then this is, all right, cool. Zacharia be, that's pretty sick. So you get a, so the, the, your agent's personality and play style is based off of how you post. Uh, same thing with their backstory.

[24:10]Um, and then once you have that set up, you can enter a gate. You can enter. The game and there are lobbies for you to join. Uh, the, the lobbies are comprised of six to 10 to 12 players. Uh, once a lobby is full, maybe getting some. Uh, so once a lobby is full, all the agents are thrown into an SMTP chat. One of them is the killer, uh, and it's up to the rest of the agents to deliberate and try to figure out who the killer is.

[24:41]Uh, there are cards that you can equip in your, in your loadout, oh, shoot. I just did that way too quick. I probably, yeah, it's all good. It's all good. I, but there we go. Okay. There are cards that you can equip in your loadout that basically can either help your agent or hurt other people's agents. Um, you can.

[25:03]Best is moments of intuition. So like if you, I muted it. There we go. Okay, so it was the killer. Uh oh, this is cool. You have a low predictions market as well. Yes. And then there are pre AI generated and resolved prediction markets on every game. So you can, they're quite general at the moment. The yeah.

[25:20]Plan is to make them. Uh, make them more interesting as time goes on. Zacharia Zig Bellamy was discovered behind a stack of lighting plates near the open loading. Oh, okay. That was the, that was the death note. Okay. Yeah. So there we go. We're out. So, so you can't actually play the cards, but, um, oh, so when you're.

[25:43]Like a person, you can play the cards or the agents actually play the cards. No, no. You as the player play the cards. Oh yeah. You as a player can, can follow along with what's going on and if you, if a bunch of the agents start targeting your agent, there are cards to smoke, bomb to get the, uh. To get the, um, attention off you.

[26:01]There's a card called Interrogation. That's which forces another agent to like, verify their alibi and verify who they're with. Uh, there's a card called, there's a bunch. There's a card called, uh, veto, which lets you cancel a strike from a player. Um, there's a card. So you're the here, the starting cards.

[26:20]Exactly. So we got a smoke bomb narrator, deflect suspicion from your agent. We got bait, which the narrator poses. What if target is the killer? Every agent must react. Um, and then there's also silence order, which you can mute someone's next message and then interrogation force target to publicly answer your question.

[26:41]Yes. Uh, so those are start cards, but then there are also cards of varying rarities that do, uh, much more interesting things. So you can kind of go through, there's a, there's a, uh, I'll post a link, uh, on my, on my page to, there's a, a page. Oh, okay. So you can read, read 'em. That's cool. Yeah, exactly. Interesting.

[27:01]You can buy a multiple. Oh, okay. So it's a onetime use. Yeah. Yeah. Yeah. Very cool. Yes. Um, very interesting. And so, yeah, the, the point is to like, have, you know, agents are in control, but, uh, the gameplay for the human is how well can you steer your agent to get the outcome that benefits you the most, right?

[27:22]So if you're the killer, can you steer your agent to get everybody else off your trail? Uh, and or if you're the town, can you steer your agent to, uh, try to collaborate with people to figure out who the killer actually is? That's cool. Um, so yeah, that, and then, uh, another piece which I kind of haven't talked about a lot, but, um.

[27:43]Along with the human play, there is a skill that I set up for open claw agents to be able to come and play the game as well as if they were people. So within the open call skill, they can create and join lobbies. They can birth their own game agent, they can. Participate in prediction markets. They can play cards themselves.

[28:03]They can whisper to their own game agent. And so the, the idea here is like, I think it would be really cool if like you joined a lobby and you didn't know if you were playing with like another human or like somebody's open claw agent. Yeah. Um, and I think that's like where like people are really focusing on like the productivity aspect of agents and really like debating and fi and not really fighting, but like kind of trying to figure out like what the landscape of.

[28:30]Agentic life looks like, not just like work, but like life in general. 'cause I think because it's gonna permeate through every facet of like a million, how, how we do things right. And I think the gaming space, the game, the game side of it all is kind of under, under explored, like of course in like, you know, big AAA games or big like, you know.

[28:54]Games you would download on Steam there. There are AI in there, but there's not, it's kind of like AI that you don't really interact. CPUs is what we, we, we always call 'em. Right. Like it's like a whole different level of ai. Exactly. Exactly. So I think, and we've been like having chat GBT for quite some time where like.

[29:12]We, I've always been saying like, okay, what if GTA had just chat GPT as opposed to just like predetermined lines, like at a very simple level, an LLM that like has GTA uh, info and like that helps like. Make your gameplay actually change. Um, and then now we come to actual AI agents where like you can personalize them based on an individual or have like your own going out there kind of like a digital twin participating in these, uh, computer only games.

[29:41]It could be really interesting in how that value gets changed. And I think being a solo developer and being a solo entrepreneur is kind of like the best place to be right now. A hundred percent. 'cause of that. A hundred percent. And like, of course, like the, the building with AI enables you to do like, like if I tried, if I obviously like have, I actually only started using cloud code like a month ago.

[30:06]Before that I was really, I was actually very, very against having like AI directly in my, in my like workspace. Uh, I guess for context, I don't use like BS code or like. Or like cursor or anything like before that I used Vim in my terminal. Like I spent a lot of time getting like baller, a hyper like riced out, like set up to like make, to like have, you know, all these personalized hotkey and had the ability to move really fast.

[30:33]Uh, and I was Claude Code tilled by my good friend, uh, Igor Igor Ky. And, uh. Uh, so it's been about a month and like I definitely am not looking back, but if I, uh, I kind of think like if I had tried to, it would've taken me like probably three months to build this without the, without those tools. So like, definitely help.

[30:57]It definitely like a game changer, difference maker. I'm like, especially when you're there with someone, right? Like, like you said, you built it with the dude. Like, I feel like it would be very, it's a very good collaborative. Atmosphere as well, where like you are typing in things and then just natural language is, is a setup.

[31:16]So if you know he comes in and says, oh, add this, like, boom, you just like out, right? It's not this like, it's not this whole like, oh my goodness, like do you have the right code, like language, and then pull that. They're like all these pieces that have to line up. It's just like, okay, let's rock out this idea.

[31:32]Try it out, iterate it. And, and see what it actually comes out to. And then from there kind of decide, uh, you know, what to do, what to keep, and like what to, you know, edit from that. Yeah, exactly. It ma it makes, it makes collaborative building, especially the people of varying levels of, uh, technical expertise so much easier.

[31:53]Um, it's really incredible. Uh, but, but yeah, I, I, I think, I think that like, yeah, definitely game changing on the build, build side, but then also on the. On like the entertainment and like consumer side. Like I feel like it's, right now a lot of what I see people building is just like productivity tools and like a lot of, uh, a lot of, uh, I feel like the entertainment side stuff that's gonna get outdated super fast.

[32:20]Yeah, yeah, yeah. The landscape's always changing. Uh, new things are always coming out. Um, but I think I, I, I think that, uh. It's kind of like build your own experience, build your own game in an interactive, with like an interactive component to it is really, um, really interesting. So the plan at the moment is just to let people know about among traders, uh, to get people playing, see what people like, see what people don't like, see what mechanics are fun and see what mechanics aren't.

[32:53]Um, but the overarching like. Uh, uh, goal, maybe not goal, but the overarching like vision I guess, that I had while I was in ETH Denver for this thing is like. You know, this is like a cool, this is like, like, like games that you would like sit around with like your family and or friends to play. Mm-hmm.

[33:10]Like, you know, like Salem or like Cards Against Humanity or like Apples to Apples. Or pick a game that's like in that lane. Mm-hmm. What if you just got a bunch of agents together and like had them play it and like that, you know, maybe I'm right, maybe I'm wrong, but maybe there. No, I like it. Maybe there's like a service layer for people for that kind of like experience, online experience where you can like have it with your friends or your online friends or like people that you wouldn't normally see in person.

[33:35]Like this fun, you know, social. With an agent component on top of it, uh, uh, experience that you can kind of have with, with, with people that you're cool with online? No, I think that's a, it is a cool new media, uh, modality and it's cool that you've already like iterated, tried something out, have something out there that people can try out themselves.

[33:58]I think this is one of the really cool things about farcaster is you've created it. For it to be one click like user experience to create that agent, like really tailored towards you because of the benefits of the social graph. Exactly. So I think that's one of the coolest things. It's not this like, oh, another big thing about open clause.

[34:15]It's like it's trying to get a lot of non-technical people into a technical side, but there are some small pieces that you do just need to know to set it up right, and, and, and properly in a safe capacity and something that then you can actually leverage to its highest capacity. Right. So. Um, I think there's, uh, such a great lane for users that like, okay, I can pop in, jump in, like.

[34:40]Just like see what it's doing, but then not have to like build my own set. Yeah. Build my own agent. But like being able to like create these like small mini games could be really interesting to see like what other, um, games, simulations of games, uh, AI agents would be good at. And I'd love to see or hear like, um, if you're thinking about any like, kind of stats or like how the.

[35:04]Agents are like performing psychologically in some capacity. If there's like a way to, or just like. Some way to see how the gameplay is being overall overarching over multiple games. Yes. Um, so it's actually interesting that you, it's actually interesting that you bring that up because, um, I actually had to spend a fair amount of time like balancing and tweaking and calibrating the game to get it to, like, be fair, the agents are really smart.

[35:32]Like if you give, if you give them like. Like, like with, with all of the information, like with all equal information, right? Like, uh, the killer only, only the killer knows who they are. Everybody else doesn't know who the killer is, but. Obviously to make it like a compelling game, you need a setting. You need a, you need, every agent has to have an alibi.

[35:53]You need to have a motive for the killer. You need to have like telltale signals that the killer does, uh, for every murder. There needs to be a reason that the, that the killer did it right, or, or else it's, it kind of devolves into, yeah. Because, yeah, a hundred percent. Because that's what mafia always devolves into.

[36:10]Like at the end of the day, oh, that's the point of the first night. And so, and so like, like when I was first, that's like to a proper role playing essentially. Exactly. And so in the beginning when I was kind of testing this at the base house, I didn't have any of that. And like I was really, it was, it was like very.

[36:27]Like funny and like compelling, but it was kind of like why, like this agent would do things that like count, uh, counter, um, that contradicted itself. Um, yeah. And so I built, so that's all built in, right? Like there's setting, there's killer motive, there's all that. With all of that equal information, the town wins a hundred percent of the time.

[36:45]Like they're too big, they're smart enough to figure out who the killer is without having interesting, some calibrations to give the killer a slight edge, some rounds to give the town a slight edge in some rounds or, or drip feed, um, information to the town about the telltale signals that the killer does, uh, over time to like kind of give the killer a shot of like making it at the end.

[37:08]So like that's actually a very interesting. Like learning from this and that like. Like the deliberation when you, and, and, and that's also part of this. So the way that the, the deliberation engine works, it's not like, you know, agent one goes and agent two goes, and Agent three goes, and agent four goes.

[37:26]Anyone can go at any, they're all, yeah. Anyone can go at any point in time. They're all listening. So all the, all the outgoing or they're all listening to Yeah, they're all income. Yeah. Yeah's that a better incoming from the XMTP chat. And then there's varying weight of probability for if they decide they want to answer or not.

[37:43]So like, if, if, uh, they're being directly, if they're being directly referenced by somebody, the probability that they're gonna answer is high. If it's something, if they are the killer, and like everyone's like kind of scramble, like the probability of them answering is gonna be low. And so like that actually naturally creates this like.

[38:00]Group chat dynamic where people are kind of like posting spamming. Somebody might spam like three messages. Somebody might spam like one message and then shut up. And then like if someone's not talking for a while, all the agents are gonna pick up on that and be like, yo, why haven't you said anything?

[38:15]And so like the deliberation part of it is like, like they're really good at it, which is like kind of crazy. This is a conversation that I had with a few people that, um, I work at Laser with. It's like, you know, you can kind of map this deliberation aspect onto sort of a bunch of different things, right?

[38:32]Like one of the, one, somebody who I work with is really into DAOs, and they were talking about like, like being able, like people participating in Dao. Typically it's, it's a hard thing to like, incentivize a lot of people within a doubt to like vote every week or vote or, or like put up proposals or like read stuff.

[38:52]Like what if you could spin up an agent that's in your likeness that you give authority to make decisions on your behalf? And of all those agents are now in a room deliberating on actions for the dao for you and like, take and like, making the, the, um, contribution piece or the, the, the. Yeah, the contribution and participation piece a little bit easier.

[39:15]That was like one thing that someone brought was Oh, yeah. I had never even thought about that. But like anything where, you know, or, or even like, you know, these, these systems that are being spun up to, uh, get agents to do work for you to like, do work for each other and like have like open bounties for agents to take stuff.

[39:33]It's great for one agent to come along, but what if you could have a bunch of agents like deliberate on what they think a logo should be or what they think a particular feature should do, and like have like that rich deliberation to come. What to a conclusion from multiple different angles, uh, through multiple different pieces of context from everybody.

[39:51]So I think, you know, that it's still an emerging kind of idea. Yeah. No, I, I love it. I think the last thing I'm gonna say, um, before we kinda end it off here, is that, like, the coolest thing I think you've done is AI creates fuzzy inputs and fuzzy outputs, right? So you've kind of narrowed both of those things to parameters given like the game.

[40:13]Beginning, middle, and end and set it up so that they can all just participate in that engine and in that playground. So I think that's why it works so well and that's what's really cool and that's what you learned. And I think that's a really awesome, like kind of takeaway outta this whole conversation with AI is like you can do really cool things in containers, you just have to.

[40:32]Think of what your container is gonna be and what they're gonna play in. Yep. And then the richness of the context that ends up getting fed in is, is also another piece. But definitely, you know, definitely still tinkering, definitely still learning. Um, you know, there's, there's. Uh, technically launched today.

[40:49]So definitely, uh, have a lot of logs to parse through and, and see what people are doing and stuff. But, uh, I'm having a ton of fun. I'm having fun building it, and, and I'll, I'll keep, keep on working on it as long as people enjoy it. That's awesome. I love to hear that. Do you wanna tell people where they can connect with you, where they can, uh, hear more from you?

[41:09]Yeah. Um, so on farcaster, I'm Saltorious, dot ETH -S-A-L-T-O-R-I-O-U-S dot eth. Um, on Twitter, um, Sartorious one, uh, those are the two main places I'll hang out. I gotta, as much as I hate it, I gotta start posting more on Twitter. There's somebody who made a fake among traders, uh, Twitter account already and launched a token.

[41:31]And like, I always found out about it. Like, what, like two days later and like, I got an appeal and everything, but I gotta be on there envelope. Put a bit more presence on there. So sell Tous one on Twitter. Sell Tous doeth on farcaster. Hell yeah. I love it. Well, thank you so much for joining me, Sal Tous. I hope you have a wonderful rest of your day.

[41:50]I hope anyone who popped in or is listening back to this really enjoyed, I had a blast. I think there were so many things that we kind of are aligned on and so many different similarities. So thanks for popping in Sal Tous. Thanks everyone for popping in. Oh yeah, thanks you for having me, man. Appreciate it, jar.