Artifact

Village Builders

alyx AI integration Check In

February 24, 2025
People
Adam Rose
Simone Torrey
James Redenbaugh
Banner
Links & Files
Summary

Meeting Summary

Purpose & Context

This meeting served to align James with Village Builder's technical development progress and plan next steps for integrating a front-end interface with their AI system. Adam demonstrated their current AI prototype, and the team discussed implementation options, timelines, and potential collaboration structure.

Current Technical Status

  • Adam has developed a basic AI prototype using the Agno framework that can maintain conversation memory
  • The system is currently console-based but ready to be put behind an API
  • They tested a first version with an off-the-shelf front-end app builder and received positive feedback

Technical Requirements

  • Create a simple web interface for the AI chat system
  • Include user feedback mechanisms (thumbs up/down on individual responses)
  • Develop access code functionality to track users
  • Ensure the system can remember previous conversations
  • Maintain consistent branding throughout the interface
  • Build with future extensibility in mind for features like relationship mapping

Integration Approach

  • James proposed using Xano for the database backend with WIZD to connect to Webflow
  • Adam mentioned they've been using Firebase but is open to alternatives for faster iteration
  • The API will need JavaScript client libraries for integration with Webflow

Team Structure & Responsibilities

  • James will coordinate the Iris Cocreative team, with Leonard as the likely lead due to his expertise in Webflow backends and custom integrations
  • Roark (based in Hawaii) may assist with product management aspects
  • Adam will continue developing the AI backend and prepare API access

Timeline & Next Steps

  1. Adam to finalize the API interface and provide access to the codebase (target: end of this week)
  2. Iris team to review the code and determine complexity level
  3. James and Roark to assemble the appropriate team by first week of March
  4. Target completion: mid to late March, with timeline flexibility depending on complexity
    • If simpler: approximately 2 weeks of development
    • If more complex: 3-4 weeks of development

Future Vision

  • Eventual development of a relationship mapping feature to visualize user connections
  • Potential for three separate AI functions: reflection, advice, and action planning
  • Exploration of integrating with a new emotion-recognition AI model (recently received a 3-month free grant)

Action Items

  • Simone to send current prototype link to James
  • Adam to provide API access and code repository information
  • Iris team to analyze technical requirements and provide a more specific timeline
  • Team to coordinate next meeting after James settles in Thailand (after March 5th)

Initiatives

AI Prototype Integration

Integrating the alyx prototype into their Webflow website using APIs and custom code.

Meeting Transcript

Adam Rose: Can you hear me?

James Redenbaugh: Yeah.

Adam Rose: Hey, how you doing?

James Redenbaugh: Hey, Rob. Good.

Adam Rose: Right on. Glad to hear it.

James Redenbaugh: How. How are you?

Adam Rose: I got really sick last night, and I think it was. So were at a friend's place for dinner and he's from Cuba originally. I mean, he didn't really grow up there, but so I. We were in Cuba over the holidays. So I brought him a cigar. Like, I bought a bunch of Cuban cigars and we smoked it. And I think that made me really sick. Like, I started getting woozy when I was there and then like two hours later it like hit me again and I was just like, oh, I feel kind of sick. And then, no, it was not pretty. And it's a thing, I guess it's called success cigar sickness. And you can you smoke too much? It goes into your stomach, I guess, and gets into your. Like too much nicotine gets into your bloodstream. It's a thing, you know.

Adam Rose: I mean, some people get woozy just from like trying one, I guess, but like, yeah, we smoke this big, like massive cigar, just the two of us. Learned my lesson.

James Redenbaugh: Well, I'm glad you're okay.

Adam Rose: Yeah, it was not pleasant, but I tested myself for Covid and flu and everything, and it's nothing like that, so. And I definitely feel better, but it was. It's a little violent there for a while anyways. How you been?

James Redenbaugh: Pretty good.

Adam Rose: Yeah?

James Redenbaugh: Yeah, can't complain. Busy building out lots of new systems in Iris, onboarding new people.

Adam Rose: Great.

James Redenbaugh: Yeah. Getting a ton done all over the place.

Adam Rose: Fantastic. Yeah, we've been pretty busy too. I've been heads down trying to get. Just working through some prototypes for the actual backend, like AI end of things, which has been pretty to really dig into and have kind of the inklings of a first version. Like we. We did one where. I don't know how familiar with any of the stuff you are, but OpenAI has something called an Assistance API where it makes it pretty easy to build like a chatbot kind of thing, but they manage everything behind the scenes. Like you can't control the chat history, you know, like, how many messages do you want to look at? What do you want to store in memory? Like, there's all these kind of things behind the scenes that they have to do to make a chatbot seem normal, like a person.

Adam Rose: Like it doesn't forget your name the next time you come back and talk to it and things like that. And so they have a way to do that, but you have no control over it. So we've been standing up Our own version of it that, you know, where we have memory and you know, you can plug in different models. So we've been playing with Claude instead of OpenAI and see if it's a better. Does Claude make a better, you know, patient, therapeutic voice, then Open AI or whatever other models we might want to play around it. So it's been really fun to dig in and I'm just putting an API on the front of it, so hopefully we can plug in a front end. Yeah, but it's been fun, exciting.

James Redenbaugh: Yeah, yeah, I'd love to see what that looks like.

Adam Rose: I mean I can show you real quick if you want a two minute demo. It's just in console right now, but let me share here.

James Redenbaugh: Two screens.

Adam Rose: This one, let me see. I could show you some. Yeah, I don't know if you want to see any of the code or not, but that's not that exciting. Oops, wrong thing. But it's just there's a framework that I'm using. There's a bunch of different frameworks for building out like agents and stuff. And I'm using this one called Agno. It's pretty new. This is the Rock project here. Yeah, I mean it's nothing major. I mean basically you just have, you know, you define an agent and what model you're using. Can you see this? Yeah, yeah. So like what model you're using, you know, descriptions, etc. Different ways to do storage and memory. So the framework handles a lot of this and then you like, you know, just have a loop, essentially get input, you know, generate a response.

Adam Rose: And then looking at, you know, there's different companies or different frameworks too to do logging and monitoring. So you want to like collect the chats and then you know the way to like train it, you have to, you want to like tag responses like this was correct, this was hallucinating or you know, provide modified examples like this would have been a more appropriate response. And so you have to have a whole like data collection and then build, building test data sets so that you know, if you change anything or change your model, you want to rerun them and make sure that it's still, you know, so people are doing like, people are creating an LLM to act like the judge or the, to decide whether the outputs were good. So you can do it at scale.

Adam Rose: Like if you now you have thousands of responses, you're not going to like necessarily want to go through them all and you know, hand correct them. So there's all these ways that people are like Scaling this up to try to, you know, basically like get it, like to roll it out in production, when all of a sudden you have like a whole bunch of people using it. You can't just, you know, like. And then you need it to be automated, essentially. So there's all these tracer, tracing, type things. And so actually the guy that I started a company with years ago started another company a few years ago called Arise, Not a creepy name for an AI company at all. They basically do this. It's like an observability platform for monitoring all kinds of machine learning stuff.

Adam Rose: And so I'm trying to plug in to their stuff and they see log all this stuff and then it shows up in their console and you can flag it and build data sets and all that kind of stuff. So, you know, just trying to get all that stuff sort of running, which is kind of cool. But the actual. Here, let me show you. I mean, it's pretty basic. Where'd it go? Horrible. Now where is the other window? Somewhere. Hold on. Here, I'll just do it down here. Yeah. Oh, okay. I literally can't see because it's all dark mode. That's the. All right, let me just make this a little bigger. Yeah. If I. I mean, so this is just that I console right now, but. Oh, what happened? Oh, I think. Hold on. I think I killed the database. I want to start.

Adam Rose: Docker doesn't want to run. Well, I don't know. That's weird. If I. Well, yeah, well, whatever. I'll just show you something here to say. Yeah. So then there's all this like debug stuff, but like, you know, you can essentially just have a conversation. It's just at a console right now, but you can see all the, like. Okay, what did I say? What did the assistant come back with? If it made tool calls. So like if for it to go and look back in the conversation history in the database it makes. It uses a tool call, which is kind of like calling a function or calling an API. So you can just see all this stuff. How many tokens did it use when it called the model? And so, you know, you can just sort of trace what is going on. Your name's Adam.

Adam Rose: You mentioned it earlier in our conversations. Would you like to share what brought you here today? So this is based on a prompt that Simone put together. We're kind of our first prototype. You know, it's not bad. We tested it with some folks and they actually said it was pretty good. We got some good feedback as like a first Cut. We did a first version where we just used this off the shelf front end app builder thing that is just sort of a, you know, I'll send.

Simone: It to you, James. I don't know. Hi guys. Sorry for being late. I don't know if you had a chance to look at it or not, but yeah, it was a plug and play front end basically.

Adam Rose: Yeah. So yeah, that's kind of. I'll probably just stop there, but that's kind of the gist of it. Oh no, I wanted to stop sharing that share. Hold on.

Simone: Great, I see I didn't miss anything.

Adam Rose: No, I was giving him a quick demo. Did it stop sharing?

James Redenbaugh: It did, yep.

Simone: It did. Yeah.

Adam Rose: Okay.

James Redenbaugh: Yeah, cool.

Adam Rose: So, yeah, I was just kind of catching James up on what we've been up to.

Simone: Yeah.

Adam Rose: And yeah, I mentioned. So I'm literally just on Friday was starting to try to put this behind an API and I looked real quick into how do you integrate custom APIs into webflow and it sounds like, I think it generally wants like JavaScript client libraries to certain. And I don't know if you've dabbled with that stuff, James, but so the thing I'm using to put an API in front of this will I think spit out JavaScript client libraries as well. So I don't know. I haven't designed exactly what the API will look like, but it'll probably be pretty basic like submit statement to and probably something I don't know, I have to figure out like does it just do that and then the response is what it said.

Adam Rose: Like it could be that simple like you know, versus you know, like calls to like oh, I want to actually look, dig through the conversation history or like more extensive things. But we don't need to do that in the front end right now. Like I think the front end would just be mostly just a chat interface I have to figure out. So you know, it's like here's what it said, here's what you're you want to say. But then we probably want to add relatively short order like feedback buttons, like thumbs up, thumbs down on either individual responses or even just a generic like feedback capture, like open text kind of feedback and then you know, make it a branding consistent as well, I think.

Simone: Yeah. So how about we pause here and let James ask some questions.

James Redenbaugh: Well, I'm still getting a sense of what's needed exactly. I don't know if I have specific questions at this point, but it.

Simone: The more talking is good.

James Redenbaugh: Yeah, yeah, that's fine.

Simone: Okay, so can I talk a little bit from like a user perspective?

James Redenbaugh: Sure.

Simone: So what we had with the other prototype, have. Have you looked at it or not at all?

James Redenbaugh: I don't, I don't see that invitation. Could you resend that? Was that.

Simone: It wasn't an invitation, it was an email that went out, you know, a while ago. But I can just screen share, basically.

Adam Rose: Thanks. Yeah, we should say that the goal right now is more to get something in front of people so they can test the interaction with the actual AI versus, like the actual app, what it's ultimately going to look like.

Simone: Yeah.

Adam Rose: Trying to test, does it actually work? Like, is this a helpful thing for people? So that's.

Simone: And one of the main things we need to figure out with prototyping is do we need three separate AIs? Because we kind of needed to do three pretty distinct things with the user. So one is reflection. The other one is once people have reflected enough and are ready to get some input, giving some advice or tips or practices on relationship building that are context specific. And so the advice engine needs to know what the reflection engine did. And then the third one is break it down into like really simple steps or even just one thing you can do tomorrow. And then it needs to remember what it did, so when the user comes back, you can kind of pick up there, how did that go? You know, and conversationally like a. Kind of like a coach would.

Simone: But we're still trying to figure out if one prompt can do. Can handle all of that or if we need to break it down into three. But we're going to start from just one. So this, you know, it. We need some kind of access code basically, so we know who used. Who used it, but just like a simple thing. And then this one just had basically, you know, a chat interface and then a couple buttons. I think we're gonna, we don't really need those, but like a, you know, just like a basic page with some explanation what it is. And then the chat interface. And then like Adam said, some direct feedback tool like, how did this conversation go? Thumbs up, thumbs down.

Simone: Or, you know, if it's easy to have something a little bit more nuanced where, like in Claude, you can give feedback if you put thumbs down. Why. Why was this a thumbs down?

Adam Rose: And I think that's on an individual. And Claude, I think it's on an individual message basis.

Simone: Yes.

Adam Rose: Yeah, because if you really try to train the AI, you want to know, like, there's general feedback, but then it's like, well, no this specific response was, you know, no good.

Simone: So how does that sound? Like, just like a simple page with the chat interface and then some way to link feedback to individual messages, basically. For now, does that sound like something that's in your. In your wheelhouse?

James Redenbaugh: Yeah, definitely. And until I can dig into testing it out, it's. It'll be hard to say if it's something that's going to be really easy or. Or really hard. And yeah, right now I can't tell because I want to say that looks really easy. But I also know when we get into things like this, we might encounter difficulties that in the past would make it really hard. Now with AI, nothing is really that hard anymore, which is great. When we encounter a roadblock, it's easy enough to. To troubleshoot and get some free expert advice from Claude on how to do exactly what we want. So. But yeah, definitely in our wheelhouse. Definitely something that we can help you figure out.

Adam Rose: Yeah. So maybe I'm wondering, James, would it make sense to. Rather than trying to scope like a definitive. Because this is kind of custom and it's not like a process, like, in terms of, okay, we're going to do branding and we're going to stand up a website, and that's probably a process that you've run a hundred times. Would you want to do a few. Just whatever you want to call it, like time and materials, like just build some hours as we just to sort of figure it out to feel, like, comfortable. Okay.

James Redenbaugh: Yeah.

Adam Rose: Do the rest of this. It's going to be what? Whatever. Right. That might kind of mitigate that.

James Redenbaugh: Yeah. Did I send you the document about our new model? Yeah. So basically, I mean, we can. We're building this model to help us manage all different scales of requests. And right now, as we're prototyping things and figuring things out, I can make it work for you guys however you want to work. So I can keep track of time and materials. But then it would be great to retrospect. I mean, the end of this month is coming up, but like at the end of March, look back and see, you know, just have a conversation with you guys about what we got done. I can, you know, show you the hours and then also put it in the language of creative units and initiatives that we're using for this new system. And I would just love your. Your feedback on that.

James Redenbaugh: But hopefully it doesn't need to be one of those bigger engagement scales. Hopefully it's more one of those support scales where we can get this done. In a couple creative units, something like that.

Simone: And going forward, James, because so probably Adam mentioned this, like, for now, we said, because, you know, some. Also looking at some other projects and what similar projects like this Rosebud, this journal, you know, they were just at the beginning, they were just using a web interface until they've created an. And so, because that's in many ways so much easier until we figured out what kind of product with which features we actually want to create, that's kind of the route we decided to go. But there are, like, certain features that we think we know we want to create. Like, for example, the relationship map, right? Like, as people talk about their relationships, how could we visualize that? You know, how could the AI feed into a database? The names, the frequency, you know, like certain, basically data points.

Simone: And then how can we visualize that in a map? And I was thinking, you know, you've thought so much about communities and relationships throughout your life, like, if that's something you'd be excited to think with us about or maybe already have some ideas.

James Redenbaugh: Yeah, definitely. Definitely. That. That really excites me. And I'm prototyping what's becoming a technology started as just like a template, but it's kind of becoming its own little tool for helping people stay connected after a gathering and giving us a unique digital space that can kind of proceed and then persist after a gathering to facilitate connections, continued connection and regathering. And so it visualizes all the people around in a circle. And then as time goes on, there's a little marker that kind of moves around the circle every year. And if we gather together again, it creates another point. We can kind of plan things out into the future. But then it also visualizes connections between people in the circle.

James Redenbaugh: So if, you know, we're all at a retreat together and then the three of us have a call, it makes a triangle between us, you know, and it's on us to go in and add the triangle. There's no AI behind it or anything like that. But over time, you see this weaving in the circle. And, you know, right now it's. It's an idea and some sketches and a part of a conversation with a few people and projects. But it's definitely along the lines of what you're talking about. And we're figuring out, you know, we're solving similar problems of how to visualize these connections with code between people. And there's a big question mark for me around how to get, how to automate AI creating something like that. But that's definitely a question I want to help answer.

Simone: Yeah. Because all these manual tools, the problem is that you just don't update it. Right. Like, forget about it or like it's just the consistency. Like the manual effort is sometimes too much, basically.

James Redenbaugh: Yeah. Yeah.

Adam Rose: Well, AI with conversational power. Right. It can ask too, which is really nice. Like, if it's like, oh, you mentioned this person, tell me a little bit about them or how long have you known them? And so you can kind of just conversationally get someone to just say it, you know, it doesn't have to be such a chore to think, oh, I got to go into this app and update this, or, you know. But I think what we need to figure out first is even what aspects of relationships we want to represent in the context of Alex and, you know, is it more event centric? Probably not, but it's like, what are the dimensions of a relationship that we think matter that we want to highlight? Right.

Simone: And yeah, like levels of trust or dimensions of trust, closeness, like how long.

Adam Rose: You'Ve known that, Are they friend, are they family, are they a work acquaintance? You know, there's just so many ways to.

Simone: Physical distance.

Adam Rose: Yeah. And so what are the overlays or ways that like, what is it that's important that we want to represent ultimately that help you know, lean in on the relationships that you really care.

Simone: About and invest and also that you need in terms of like, you know, when we're thinking of like mutual aid type of engagements, we got to figure.

Adam Rose: That part out, obviously. And then I think it's like, okay, well now what are different ways to visualize that? Right. And ultimately with the goal of like, does it help them actually take action? Right. Is this a helpful aid that is like, oh, this really helps me see where I'm like, I really should be investing more time over here, like in maintaining my existing friendships or do I really need to find new relationships? Are there gaps in this map? You know, so we need to figure out, like, what are the questions that we're actually trying to help answer by surfacing this information so that it's helpful and that's more of a, you know. Yeah, product exploration.

Adam Rose: So I don't know whether how much of that is going to come out from people testing this and then we get a real sense of it and then we make. Or do we need to actually really experiment with visualizations? You know, I don't know how much we're going to need to like, you know, Learn into that, I guess is really.

Simone: But like, currently I'll. I'll send you our news pitch deck. But currently like the four main features that we're thinking about is like, so one we're now calling like a journal, basically like a relationship journal. But it can also be. We just don't want to give it a human character. We don't want to call it a coach, but it is a conversational back and forth where you're surfacing your challenges or your joys or, you know, you could like process your relationships and then it stores data from those relationships that can like feed into the map or that can like help surface certain practices, whether it's. And that's the next feature, a practice library that kind offers relationship practices from like frameworks from authors that we partner with or nervous system regulation practices.

Simone: And then the third one is that like basically the challenge of behavior change. Right. Like, how do you really get people from like having learned a new thing to actually doing that in their lives? And we call that the social navigation system or something. You know, it actually helps you like get there. Those are all draft names, but like breaking it down into small manageable steps and then like working with like reminders versus calendar, you know, like all those things that help us. Or like even gamification. What elements of gamification do we want to build in? So that's also a whole thing we need to learn into.

James Redenbaugh: Awesome. Awesome. So there's training and designing the AI to populate a database and then presenting the content of that database in a dynamic way. Yes, in a very dynamic way. Yes.

Adam Rose: That's definitely a bit later. Like right now it's really just. We just need a front end that we can.

Simone: Yeah.

Adam Rose: People to.

Simone: But I just wanted to, I wanted to map out a little bit the scope of what we would be excited to work with you on because there's so much experiential alignment. You know, you just know so much about communities and relationships and.

James Redenbaugh: Yeah, yeah, awesome. And we def. You know, whatever we build now, we want to build in a way so it's easy to keep building on it.

Adam Rose: Yeah, I, I haven't looked into this at all. Like, does webflow have like. It must have pretty decent mobile support too. Like, but are, do people. Are people building like mobile apps? With Webflow.

James Redenbaugh: You can. Because you can export your Webflow site as just the JavaScript HTML CSS that it's built on and then turn that into a mobile app. But it's. There's no easy way to do that. And, and then your website and your app would be two different things. You can have a webflow site that pulls from the same database that an app does.

Adam Rose: Sure.

James Redenbaugh: So you can build an app that works the same way that the website does in there, connected to the same thing with the same kind of auth and user access. Yeah. And so for all this, right now I'm thinking about using these tools. One is called Xano. Have you heard of Xano?

Adam Rose: That sounds familiar. Xano, Zano and Bubble. I, I'm not familiar with it enough to know exactly what you're talking about.

James Redenbaugh: But yeah, check it out a bit. It's a, a really powerful tool for building databases that are easy to connect to and integrate. So we could use Xano as the back end, essentially use different APIs to connect the AI to Xano and then use WIZD or something else to connect Webflow with ZNO. So the Webflow site is the interface, the AI is the intelligence and xenos where all the data is organized and stored. And then Wiz can take that data and serve it to the website in whatever way is needed. So if it's a chat response, it shows up right there. If it's a relationship map, it takes that data and runs it through a JavaScript to present that in a certain way. But it's all right there immediately. You don't have to refresh the page, reload anything.

Adam Rose: Let me take a look at this. So so far we'd been using Firebase, which is Google's kind of app dev stack that provides a lot out of the box. But again, all we've done with it is some very light prototyping and an early version of the app. So I think I've had it in my head that's what were doing, but it doesn't necessarily. Again, it's not like we've got a big investment in a whole bunch of app development. It was just like some stuff. So I think to me this is more about rapid prototyping. What's quick, that we can just get stuff up, stood up quickly.

Adam Rose: And if this is what your team knows how to do and we can iterate fast and owning the whole kind of app stack instead of, oh, I just own this one page that, you know, I have to wait for you, Adam, to provide me APIs to try to like talk to. Right. Like, this might be a much easier and faster way to iterate. So I want to be very, I hear you on that like that this might be just a faster way to do stuff. Yeah.

James Redenbaugh: And I'm not saying like this is totally going to be the way to go. It's just my thinking right now and I definitely don't want to make any decision about that immediately. Yeah.

Adam Rose: I look at it because it does look interesting. And you know, Google is interesting and everything too, but it's also got its own challenge. Like I've, you know, it's been hard to navigate the Google ecosystem a lot and so I find a lot of these smaller things are just. There's a lot more streamlined to like, this is what I want to do and it just does it well. And Google's like.

James Redenbaugh: Yeah.

Adam Rose: Or like bigger scale companies.

James Redenbaugh: Yeah. And we have a guy on our team who's been doing a lot of interesting stuff with Xano and Wizd on webflow and he's helping us build functionalities like this out on our own website using these tools. He's offline in Colombia though for the next two weeks anyway. So he won't. He won't be much help immediately.

Adam Rose: In the meantime, simple thing like with the web page I'm talking about and not, yeah, stuff but like otherwise it's like, okay, just call this API. You have a key. Just use our chat kind of thing. But if we want to start.

Simone: Time check. Yeah, it's 10 pass. Do you have to run?

James Redenbaugh: I don't think. No, I don't have a heart. Stop.

Simone: Okay.

James Redenbaugh: Do you guys have to run?

Simone: I have one more thing that I want to introduce to you that we found out about. And Adam, what about you? Are you still going to run pretty quick?

Adam Rose: Yeah, I have a 1 o'clock.

Simone: You're still going to that?

Adam Rose: Yeah, I'm actually feeling a lot better and I think I figured out I was telling James that I got really sick. I think it's literally from we're at Gus and Lila's for dinner and we had a cigar and I think there's this thing called cigar sickness that you can get.

Simone: Oh.

Adam Rose: From smoking too much. Like it can. Literally all the nicotine can get into your blood system and it's like.

Simone: Okay.

Adam Rose: A thing. Anyway, I'll.

Simone: I'll check with Sean then about the. In person if she still prefers to do.

Adam Rose: But if you want to check and I totally get it if she's not comfortable meeting.

Simone: Okay.

Adam Rose: All right. Do you want to stay on? I do need to run. I have to drop Ana somewhere on my way too.

Simone: Okay.

Adam Rose: Yeah. But yeah, let's. I'LL look at this, we'll definitely pick it up. And she's gone.

Simone: Can we, can we reinstate that meeting we had next? Oh, was it this Friday or is that too soon? That was this Friday, right.

James Redenbaugh: I can't do this Friday. Okay.

Simone: Oh, yeah. You're going to Thailand, right?

James Redenbaugh: Yeah, I'm going to be in Turkey actually until the 4th and then I'll be in Thailand.

Adam Rose: Okay. This summer too.

James Redenbaugh: Oh, fun. Yeah. But we can find a time when I'm in Thailand. I can send you my. What works for me. I'll be getting up early, taking calls in the morning there.

Simone: Okay.

James Redenbaugh: But Adam, if you want to send me what you guys.

Adam Rose: What.

James Redenbaugh: You guys have so I could start poking around and if Simone, you could resend me that demo that I couldn't find, then I can start announcing.

Simone: Yeah, yeah.

James Redenbaugh: I'll start getting a lay of the land and have a much better sense of next steps.

Adam Rose: Yeah, let me try getting this to the point where it's. Because it'll be easier to act like where it's behind an API and if, then if you just want to play around with what's there and that might be easier. I don't have any of this stuff put into GitHub yet, but I could do that too if you want to look at any of the code.

James Redenbaugh: Cool. Okay.

Adam Rose: All right. Great to see you, James.

James Redenbaugh: Good to see you, Adam.

Adam Rose: All right.

Simone: Yeah, so just a couple more minutes, James.

Adam Rose: Yeah, I'm going to go.

Simone: Bye, Adam. So one very exciting thing we just found out about is this new AI model that is trained on recognizing human emotions. So it's like specifically trained to be a more humane AI that's more like attuned to who's talking to it and how they're feeling. And so it's like basically made for our product, but currently it's all voice based or even like video based. And so that would be a different product. They're working on a voice on a text to voice, but obviously it's a lot easier to recognize, you know, how somebody's feeling from their tone. Like there's just a lot more information. But in one of the next iterations we want to start looking into how it would be to use this. And they're basically like a B2B version.

Simone: And so their go to market strategy is that they're giving free grants to startups to build their product with it so you can use their thing for free for three months. And we got that grant basically. So.

James Redenbaugh: Nice.

Simone: Yeah.

James Redenbaugh: Awesome. Yeah, that's exciting.

Simone: Very exciting.

James Redenbaugh: Cool.

Simone: This whole market is moving so fast. It's so crazy, you know, like, two years ago, these AI models, they couldn't even do memory, you know, they couldn't remember anything. And now it's like. Yeah, it's insane.

James Redenbaugh: Yeah. Yeah. Wow. I think you guys are in the right place at the right time.

Simone: I hope so. It feels that way, you know, it feels like every step then is like, oh, and here's this thing.

James Redenbaugh: Wow.

Simone: Amazing.

James Redenbaugh: Nice. Cosmos is conspiring.

Simone: I hope so. Yeah. What are you going to do in Thailand and Turkey?

James Redenbaugh: We're taking my dad to see my sister. She lives in Thailand and hang out with her. And in Turkey, we're. We're just. The four of us are exploring and having fun.

Simone: Who's the four of you?

James Redenbaugh: Me and my partner and my dad and his friend Stella.

Simone: Oh, cool.

James Redenbaugh: Yeah, yeah, it's just fun, quirky, long family trip that I'll also kind of be working on and. Yeah, it'll be great. We're looking forward to it.

Simone: Awesome.

James Redenbaugh: I haven't seen my sister in a while, so it'll be good to hang out with her for a bit. And then were planning some, like, special day trips here and there over the three weeks to see some cool stuff and get some R and R. Awesome.

Simone: That sounds sweet.

James Redenbaugh: Thank you.

Simone: So would you be working on this with us then, or would Ivan or some other team members?

James Redenbaugh: Yeah, it sounds like it's up Leonard's alley. Leonard is this. This German guy. Who's Leonard? Leonard. Yeah. He's more. He's super competent in webflow backends and custom integrations and stuff. I definitely be involved. Avon's more, you know, simpler front end stuff, so we'll likely still use him, but the complex integrations are going to be a bit over his head. Yeah. And I don't think you've met Roark yet. Roark is a great team member that we have as well. He lives in Hawaii, but he works on pst. He gets up early and he's doing a lot of product management with us, but he also has a lot of coding experience and webflow experience, so he might be of assistance here as well.

Simone: Cool. And what. What do you think is our timeline? Like, when. When can we. And how do we. How do we, like, kick this off?

James Redenbaugh: Yeah, good question. So I want to dig into the code a bit and wrap my head around it a bit more, but.

Simone: Okay.

James Redenbaugh: I think it sounds like the sooner the better for you guys getting something. Getting something workable. So I definitely want to. Want to get something working this. This coming month. How soon in the month will depend on how hard it's going to be.

Simone: Yeah.

James Redenbaugh: And if it's as easy as I think it is, you know, mid. Mid month is a good target. If it's not as easy as I think it is, then at three or four weeks, something like that.

Simone: Okay.

James Redenbaugh: But hard to say. And also if. If you guys want it sooner than that, you know, we can bring in more of Roark's time and attention to. To make it happen. And. And he can even leverage other. Other experts that we know to. To bring them in and make anything happen.

Simone: And so I can send you the st. Like the, you know, kind of the basic thing right away. And then Adam said he needs a little more time to, like. I don't know what that means, but, like, hopefully by end of this week, and then you would dig into the code and. And then, like, put together a team or something by next week. So that's the first week of March. And then it'll take you, like, one or two weeks to put things. Or one to three. No. What? Like, two to four weeks, depending on how hard it is.

James Redenbaugh: Once the team's in place and we have a clear plan, might be a week, might be three.

Simone: Okay. Okay.

James Redenbaugh: And once we really know what we're doing, will have a clear sense of how long it's going to take to make it happen.

Simone: And how long do you think it'll take us to figure that out? And is that mostly a you guys thing, or is that us in conversation? We'll need a few calls.

James Redenbaugh: I think mostly an us thing. Once Adam sends me the access to the code stuff that we can dig in.

Simone: Okay.

James Redenbaugh: Cool.

Simone: Yeah, I mean, there's not really, like, two weeks more doesn't make a huge difference. It makes a little bit of a difference because, you know, every time we push the timeline, our fundraising ability gets pushed a little bit. But I think mid March is. Is a mid. Mid to late March is a really good target to have a next thing out. All right.

James Redenbaugh: Okie dokie.

Simone: Okay.

Adam Rose: Great.

James Redenbaugh: Exciting stuff.

Simone: Yeah, super exciting stuff. And, you know, we have funding for, like, the next three months or so, and then we're. Our next goal is to do a friends and family round. So, like, let us know what. How we can best work together to, like, get your attention on this.

James Redenbaugh: Okay. Okay, sounds good. Sounds good.

Simone: Okay.

James Redenbaugh: Yep.

Simone: Great.

James Redenbaugh: All right, I'll look for that stuff and follow up also before I take off on next steps. And great talk to you soon, okay?

Simone: Yay. Bye, friend.

James Redenbaugh: Good to see you, Simone. Take care.

Simone: You too. Bye.