# 🐣│cohort-voice
- Session: 0f93e4bf-fff8-44e0-86a3-370981a39194
- Channel: Discord #🐣│cohort-voice
- Started: 2026-06-24T16:58:48.015Z
- Ended: 2026-06-24T17:40:37.337Z
- Participants: takekek, duckanbro, graven | Flow State, Aphilos • Pharo, ECWireless, samkuhlmann
[00:00:00] [duckanbro]: Okay, so uh thanks for joining me then uh we're doing just kinda like these um these quick interviews, we're trying to keep them to 30 minutes time and and uh science side chats and maybe just trying to get the different perspectives on how people are using AI, how it's changing the field they work in, um how to adapting some things like that. But we're quite we've done um I don't know, maybe six or seven of them so far and it's really interesting just hearing different perspectives. Um we've got people from like from developers, project managers to um like an SEO person and content creators and product builders and and just uh advertising authors. So we're just trying to like just see how people are are using it. And it and sometimes it we maybe just sort of f I use judging between sometimes it's like uh finding this whole infrastructure and and a local system or whatever it is and also so there are just kind of some basic questions um and we just trying to catch other people's perspective. And what we do with the recording is we f summarize it and kind of turn through it, set out some some content, some blog posts, some like YouTube cards, some deep dive like further research opportunities. But uh and uh to be done a lot of it is pretty sloppy thing, uh but it's giving us uh surface area to test a lot of stuff, and also it's helping us inform our strategy um as an agency. So that's kind of our position. Yeah, um it's it is public and each session has a little bit of content attached to it. I think it'll become um really interesting when we kind of get through this month and then kind of start synthesizing all the things together and looking through the few lines and um finding some new interesting kind of things to to deep dive on. Um but I'll definitely share that with you. And uh as far as if it's public we're we are recording, but um drawing things up on uh great field YouTube. If you don't want the the interview recording up there that's fine. Um just let me know. Otherwise it's just like summaries and stuff. Cool. Well, first off, thanks for joining um maybe we could start with a little introduction. Yeah, interesting. So you mentioned uh you know being a product manager. Um how has uh that changed uh you know uh you mentioned a bit if you're you're getting more into some more dev stuff.
[00:00:00] [graven-Flow-State]: Yeah. Yeah, makes sense. Uh people trying to talk through what I've seen and hopefully I I guess like yeah, I'm always looking at the road, so I'll I'll check out whatever sort of content doesn't put down my are you are you putting off that stuff publicly? Like you said? Yeah, which is kind of fun. For sure. Uh yeah, no, mostly they're uh company that does for long time now. Uh been proposed since full-time since 2019, always been an independent builder focused on kind of public goods open source, uh kind of radical market type ideas. I always focus on kind of just exploring and doing funky fun stuff. Um I am personally more of like kind of a product manager or something along the online. I'm not a developer, but I have been working with developers basically my entire career, kind of managing that and so um technical-ish. Um and I guess we'll probably get into how that's changed with with using AI, what that means for like my role and what I've done. Um recently it's working out full sync with the extremely funding. Um the superfold, um trying to use programmable and those streams to move meter around more effectively, focus on public goods. I'm kind of actually this is uh we were partnered with Octant and do a lot of stuff with them, they're a bit public goods vendor. More recently I've taken on a full-time role with Octin's, focused on kind of uh well, kind of just getting started there, but um helping blow out the V2 protocol, making builders, um Ts and um trying to uh create uh a community of builders around that um that protocol. Yeah, and it's it's evolved, yeah. I guess maybe it's since the last move when it really started taking off for me with kind of actual like kind of proper famous agents sort of stuff with quad code. Um this we like started off trying to build uh like a whole bunch of stuff as I even just like fucking around finding out and see see the see what happens. I think it's settled into uh less uh less aggressive like me trying to like I'm still building stuff and shipping code in a way that uh never would have been possible for me unless I sat down and taught myself like 70 or whatever. Um so that's continued, but I think I've settled into a way where it's like I still
[00:00:13] [Aphilos-Pharo]: It's quite
[00:03:00] [duckanbro]: Uh how has it changed with AIDN? Yeah, that seems like a really good uh a great use case is like you're doing job and really understanding what it does and um downloading my history a little quicker instead of spending months at it just to try to synthesize something. There's a few things I wanted to dive into your sooner so it sounds like you may be using code um are you using the kind of in the context of the cloud code um are you know things like cursor or um VS Code innovations or or things like open car or f or anything like that? Are you using uh GitHub control pilot? Yeah. I would like to make life is GitHub or Git in general. There's just uh it's kind of one of those tools that always sits outside everyone knows in circuits how to use it, but it's actually pretty powerful once you give it to a uh agent, then you might just change your work trees and and um managing mergers and and it's it's pretty interesting, but anyways. Um has just like a deep creative integration with GitHub, which is sometimes nice. Um anyways, dead merging. The other thing that I thought was interesting in the product manager and product owner and person. I mean, one of the main things is you're working with a lot of people with diverse skills um across the design products all the time. And um so yeah, I was just wondering about those relationships. How are how is it working with other engineers that are also using AI like uh how's it how does that change? How do you collaborate and coordinate on things? Yeah. I mean uh adapting the process um is very interesting. Have you had issues with extra creep you know like everything's impossible now, so let's just do everything. Or um have you met that you've had to add a little stronger on and like no we gotta get the structure for now? You mentioned uh uh well uh quickly just uh you mentioned some skills and um the processes and some new uh finance uh
[00:03:00] [graven-Flow-State]: like working with like more that's work with some of you even before like smaller teams um find an opportunity define like uh how we're gonna attack it create a design and iterate through with like a single developer a very small team to to go try to build that um ai has changed that in terms of the seed of the maybe the ambition of what we can take on. I don't think that's moved to every everyone here. And I I like really I contribute some code to that and I can like feel good about like maybe being a less annoying as like a perfectionist or something like that. Where I can like if something just like slightly needs to change like accountability can can make those changes. And personal projects and like in exploration I'll take something from start and finish. But instead of going to like sort of working with like a technical partner it's going to have a little bit more like depth of understanding and help uh architect the system in a way that's not just like I don't know not going to fall into a trap of um thinking I is production ready or uh something that like one shot is is now like ready to circle like funds and contract funds or anything like that. Um so uh yeah I guess and in my role with like Octins um they've been doing a lot of work uh I mean researching the code bases using that as like using um code yeah like in the context there of like a GitHub remote to really dive in deeper and expand mind like one catch up on things that uh reports prior to being around um and really interrogate like what what's going on, how does this whole system work, um building a lot of diagrams and things like that um using using um cloud code. Yeah. Um I with when the first hype cycle of the claw came I I set one up and then mess around a little bit on that but haven't been really used that my day day to day work. My my user setup is Squad Code, my terminal um I keep kind of GitHub desktop side by side with it um and uh just kind of work through that. I do have like I can I guess like outside I it still work stuff but then I'd really like oh my mobile use I use cloud
[00:06:00] [duckanbro]: Yeah. I think it I'm one of the interesting things that I almost it surfaces some of the problems that happen in the code that you might not have cared about. Because it's looks like a literal interpretation in the back end sometimes. And so you're like, oh, can I make a a view of this record from the database and it's like, oh my god, why is there so much stuff in my data model? Uh anyways, sometimes I'm data for that. Um are you using uh like some of the features of hard quality because uh things like tasks and automation workflows that could be found use cases to like run something daily or product metrics or things like that. So what other things are you still a little hesitant with the uh those open besides front uh interface design harder I know you're working a lot with smart contracts you have, um are you are you starting to get higher confidence as as we go f um one of the things you were mentioning that with the skills are doing are kind of um second skills and um and we talked about magic skills. One of the things I that I've noticed from some of the other toxic analogies in this case is that the target can ship. So and you're working directly with um what tips have you found to like to actually uh get something to ship production ready so you're not just you know dealing with mountains of slot or internet f new features and and things like that. Yeah. I mean you kinda get I think this is with anything. There's always been kind of like uh like uh flood flood everything into as much content as possible and read out to microvalidation that way. Or do uh super quality content uh up front. And so it's kinda like why the quality kind of slipped there and the quality side is so like very takes time, although we have some shortcuts that there's some things you just can't sit, you know. Yeah, um congratulations on the new role. Um what do you do? Excellent. Uh Sam's learner.
[00:06:00] [graven-Flow-State]: Typically um for like the minimum chatbots sort of stuff. Um I guess I could say I no GitHub code pilot today. So um I do have then also like I've experimented with with Codecs uh a little bit and I have like an open AI subscription um big that the literature so uh I don't know if there's something that like I want to kind of compare and contrast and use those those two. Um I've tried and I in the past never really liked any of the results or suck. It's never felt like it uh I don't know. Even when the like I I can't even remember the model numbers, but like Gemini whatever came out and it was the biggest status model. Um I tried using it for some stuff and it was just like a really it's maybe it had all the parameters in the world, but see they would have the context awareness that made it useful with strings. Yeah. Yeah, like I'd still say in like my I worked with uh my partner Stephano for a long time and so it was like a deep relationship, but I think we've always kind of worked maybe differently than like uh I mean you work on a sprint or like doing a a whole bunch of like like kind of project management stuff where it's with CP where you can kind of mind meld in a different way and keep it right. And that's always been my bias. Like I've done the project project product manager stuff, but then in all these other ventures, I've always had to do everything else too. So um I guess yeah, I've always kind of uh had uh like a hands-off approach to kind of like when managing and working with uh the developers that I work with. Um I'd say earlier it's definitely changed the soothe iteration and um the ambition of what we can do. So in a case so I guess uh it changed the bottleneck of what what we were doing. All of a sudden, like when we were shipping that's step modus creates uh new features that you know take like a week and you knock that out overnight. Um I was also the main tester. And so I ended up my workload of like testing app especially like I mean when you ship fast and like you feel about it, but like there's those rough edges and missing pieces and like the business logic failure isn't there. So I ended up like
[00:09:00] [duckanbro]: Well that's it's a better answer than most of the Oh what's going on? Uh one one last question, we're almost out of time, but um do you think like so what's your kind of uh take um where uh product development and and uh and other things that you're doing, but yeah in general, what do you think it's gonna be in six months or a year? Do you think there's optimism in that or is it pretty doom? Yeah, that's uh scary side. Well thanks, Draven. Uh thanks for hopping on and doing a quick chat with us. Uh if you have um anything you want to share, uh links to closer and faster for your ex. Um I can try to dig emo if you got links that would be helpful. Um we uh three minutes if you gotta drop two cents. Um otherwise the most of us would probably just continue chatting or feeling in. Cool. Taylor said uh how are you feeling about this air compared to past ones? So crypto question. Yeah. Um
[00:09:00] [graven-Flow-State]: I was sending like all day testing as like just a a tester to just make sure that the product that went out the door was like that I envisioned and maybe that was like a result of not secondary enough up front and you know there's there's trade-offs there. Um what we ended up doing with that process, like started saying, well what can just do for us to like lose up the bottleneck on on testing and documentation and all those sorts of things. So we only started developing the scores and kind of like our own harness and stuff to to help uh alleviate kind of different bottlenecks that you're now like the the shipping process, um, so that like that my time wasn't just over one being a tester. Um and uh I I think that process like just like taking a step back and acknowledging that the bottlenecks in software development have completely changed. Umgoing one and like changes what like their value is created in that pipeline. And then when you can ask it, because if everyone can throw their five again or whatever, um like then that becomes kind of table so we're all that we for sort of uh so many people that are on the edge here, but um did ship changes like all of a sudden your tastes, your like decision making on the where the the laptop obviously then more important distribution, and those are the sorts of things that I'm kind of uh thinking to now as like where we all go. Yeah. Um just thinking back to early in the year, like when this community started taking off. So we're actually in a position where like we have like a a customer that wanted some new new features and we're kinda like slowly, like it could we didn't really know, we hadn't really like internalized and so maybe it was really cautious there, but like all of a sudden we could add all the things that they asked for, and it was actually the necessary things, things that like were on a deadline that would have been impossible previously, but a lot of us to they said he jumped us on and said, Yeah, sure, let's do it. Um that then smoke too far the the other direction where any time we have a client is just like, look, could you do this and could you do this? And if the answer is always maybe yes we could, but should be, um we definitely had to like dial that back a little bit and just say like we we we what's what's the advantage here? Like what are you actually gaining uh through this? Um so um
[00:12:00] [graven-Flow-State]: Yeah, so I guess we we d like kind of through the pipeline did like a uh spec spec writing, tech eval, uh like feasibility, like and then just built out the scale for each each of those steps, or you know, when I say me as loosely we use like I was working with Stefano to kind of define those things and he definitely pushed that for. Um and so that was kind of like tailored to things how we work the the tools that we have, the the fulfill sort of the things that we were focused on stylistically, you know um and all that sort of stuff. Um that's continued to to evolve. Um I guess also uh I don't know for sure, like I I love the the Excalibur MCP scale, like I use that like the MCP all the time. Um like kind of diagramming and like that sort of thing. It's just like it I I I was a big Excalibur user before. Um that's where I just mop like in my multi-hack where I would just like create uh uh wireframes in Excel draw. Um I don't really get wireframe anymore, um, but uh definitely I ran in is the number. I have tried cloud cloud design um a little bit. Um hasn't hasn't necessarily like one day or anything like that. Um I definitely actually like design and front end stuff is like the most um base part of kind of like building that doesn't feel like it it it doesn't always wow me. Um yes you can create but it's it's gonna have rough edges and it's just kind of like I we all know that it if you don't uh come up with data design every look like every other website out there and it's like yeah a soft website and we saw it from the f no I I said Yeah Yeah it I mean it's definitely grown and now more and more like the the feature that gets talked about most is like yeah how good uh these bigger models are at security uh so um I haven't shipped any smart contracts with uh like that's uh uh messed around and built built some but I just like I haven't gotten to the level where I would like say yeah like shift as as something that I built. As I've worked with like obviously like different Doctor doll like Stefano and Sam. I mean like uh s I I know like
[00:15:00] [graven-Flow-State]: using these tools to build smart contracts. But I it it's almost like I I would want so like a professional developer to to take a look. And maybe like at this point like the models or favor a little bit back access to that is gonna be like obviously beyond the scale of of near motives. Um but I would still kind of like I don't know there's still I guess I'm just not there yet in terms of saying like yeah apply code with a smart contract and let's roll it out. Maybe something dead simple um it's would be fine when you start there and build up to something that's more like of a system rather than just kind of like a single one function or whatever. Um but yeah we built up that total confidence yeah. Yeah I guess um we've always been pretty iterative I think um and I think that helps a lot. Um a lot of the work in the last I'd I'd say like six, seven months um there's not gonna be like a base there of something that we were building off of it's kind of like making enhancements um for a client. Um with Octa there I guess there is a little bit like it's saying like I'm building off I'm building things off a base of the the seemed already built with that service like kind of these different frameworks or like allocation mechanisms. And so we I most you could maybe describe it as doing a lot more work from like one to a hundred than zero to one. Um zero to one framework, I guess what we've done we wouldn't shift any of these as like a production things but like going from an idea and setting up an MDP and like time boxing that to like a burger two and getting something that you can click around on and feel and kind of validate the idea is like super powerful and like um s I look at that not as like a burden to like try to ring in this but like previously like I guess we would maybe like debated and thought through and scoped and spent like way way more time for like validating knowing that like the cost of actually like trying to to ship that thing is orders a magnitude more. So I guess I'd kind of like it was leaned into like yeah you can do a sloppy search version as like an intro and demo like validation and then work from there in terms of like well this deserved to go
[00:18:00] [graven-Flow-State]: So I yeah, just linear data relation for for me has always been like the best model. Yeah. Well like I think there's a lot of work with AI like a barbell it creates a barbell effect in like multiple ways. And like h how you approach your strategies and like on the low end, throw a bunch of shit at the wall and not worry about it too much in terms of kind of like that as a process and then going like hyper handcrafted and like once you find your winner or something that you you want to to go off or after. Um and then I think that like changes kind of like structures of companies too and like hyper you know, super cheaper building companies and super small ones, um or teams or um whatever sort of organization we're creating. The core dev um uh building relations. Um I guess it's um yeah, like trying to try to build a uh I've been around the space kind of just like always bottoms up building, like some of the good funding sort of stuff. So I'm knowing with a lot of the kind of target audience, people that they're able to try to sleep and build on top of kind of the vets and other kind of pieces of protocol that are wearing out. Um so helping craft a strategy to like turn that protocol into something that people can build on and create like kind of sustainable businesses or uh or operations around and not just rely on grants. Um and so yeah, it's starting pretty recently. Um so have been kind of getting up to see like a few more things. But we were working with them before, so it wasn't starting totally from zero, but um hopefully be wearing out some of the the the s the stuff from this role on the impact and engagement um person. I I was just gonna respond in the in text and um that would talk talking and typing, but obviously not they just have a dump salary, which like drew from the so uh yes we do have the nuggets number, but uh I don't know, maybe our days are numbered now that we're dumping salaries and uh to have a trade graphics to do it. Um it'd be boring to say that France they work there. Um I originally mixed before before the the and that that's not like a uh
[00:21:00] [graven-Flow-State]: It's way out there either. But um I did that very prior. Um little both. I would be like there's depends on which side of the bed I work up on. Um in some ways like the market is kind of coming to me and what I've always like promoting as much of like working with like a small team and and and like not doing the politics of like a larger organization and like just the overhead that comes with trying to coordinate a lot of people. Um and so uh in some ways like you you very clearly can say, hey, like if you have a small team that kind of knows the direction and sort of work together, like this is clearly like supercharging for that. Um on the other side, I also see that like uh as soon as like just development is a bottleneck and it pushes it to other parts, like distribution becomes more and more uh valuable and like existing distribution channels and like owning customer eyes and advertising, a whole lot of things that I've never like never uh actively kind of pushed against and not really uh in my vibe. Um so I guess it's again that barbell sort of thing. It's like there'll be opportunities for smaller teams that can move quickly, have different cost structures and build stuff. Um and that's really uh uh exciting for someone that's entrepreneurial, um, and has kind of been doing that. Um but there's obviously gonna be a lot more people and it seems like big players are also gonna do in a lot and that creates um I don't know. That is definitely not as exciting. Uh we'll hang and start dropping spokes. It's a bit hidden harder, uh left and right, like I guess um the only like uh excitement in the world of like the fix or tech is all AI related, and so it's like 90% that and crypto maybe takes a back seat, and yeah, there's uh XLR2 and like things uh story to be told for agents using like crypto Rails. It doesn't feel as uh tangible in like a here now for like a consumer level. It feels like everything's kind of uh so institutional stuff, and maybe as a result of that adoption, there's opportunities for um bottoms up builders and like different opportunities, but yeah, definitely doesn't this one feels harder than any other smaller.
[00:24:00] [graven-Flow-State]: I believe so.
[00:24:02] [samkuhlmann chat]: Did the timberwolves utilize AI to beat the Nuggets in rd of the playoffs? (https://discord.com/channels/684227450204323876/873247103231610931/1519392374122348755)
[00:25:37] [samkuhlmann chat]: <@654810759951417344> ☝️ (https://discord.com/channels/684227450204323876/873247103231610931/1519392772082368712)
[00:29:54] [samkuhlmann chat]: don't need ai , when you have ae (https://discord.com/channels/684227450204323876/873247103231610931/1519393850500710591)
[00:33:05] [takekek chat]: how are you feeling about this bear compared to past ones? is crypto so over (https://discord.com/channels/684227450204323876/873247103231610931/1519394647917592716)
[00:33:05] [graven chat]: https://flowstate.network/
[embed] Flow State - Streaming Funding Solutions - Continuous funding apps, payment tools, & incentive systems powered by Superfluid: Flow Councils, Flow Splitters, Flow QF, & more. - https://flowstate.network/ (https://discord.com/channels/684227450204323876/873247103231610931/1519394650350157834)
[00:33:16] [graven chat]: https://x.com/GravenPrest
[embed] graven (@GravenPrest) on X - dev & builder relations @OctantApp\. advisor @flowstatecoop @Superfluid\_HQ\.
I don't have a content strategy\. - https://twitter.com/GravenPrest (https://discord.com/channels/684227450204323876/873247103231610931/1519394697880010773)
[00:34:36] [graven chat]: https://octant.app/dashboard/home (https://discord.com/channels/684227450204323876/873247103231610931/1519395030614413362)