Matthew: Welcome to the Restock Podcast by No Code Supply Co, where we highlight people, the things they build, useful tools, visual inspiration, and conversations to refill our shelves and ourselves. Hey everybody, I'm Matthew. Corey Moen: And I'm Corey and today we have the legendary Timothy Ricks. So stoked to Timothy Ricks: Hey everyone. Matthew: Legend. The wizard. Corey Moen: have you. Thanks so much for coming. Especially anybody from the Webflow space certainly knows Timothy. and yeah, man, thank you again for coming and sharing what you've been up to and how things are going. Timothy Ricks: Thanks, so glad to be here with you guys. Corey Moen: I think we were talking a little bit ago on just, yeah, some of the some of the ways that you've been building personal projects. And of course, you know, I'm sure a lot of people are excited to learn how it's going since you joined the team at Webflow. yeah, how's it been? Timothy Ricks: Yeah, it's been really great. I'm still going through onboarding and stuff, but I've been able to, you know, give feedback and like see how the team works. And it was a big eye opener coming from agency life to in-house, just like the the amount of things going on behind the scenes. It's it's really a big eye opener. So it's been really cool. Corey Moen: Mm-hmm. Matthew: Yeah, it it's a it's a whirlwind, it's a blur, huh, for sure. Corey Moen: Yeah. Timothy Ricks: Yeah. Matthew: yeah. So it's pretty exciting because Timothy is is on the same team as me, which is really exciting. So we are both forward deployed engineers, and which means that we go and we embed with customers, with kind of enterprise clients directly and help solve a situation for them, actually get hands on keyboards and actually build solutions and hand it off to them. So we're still kind of in the early stages of forming this team. At Webflow. so it's kind of exciting to be founding members of this team and really explore this new area. And we are focused on source. Our team is wholly centered around source. making sure that it is going to become the best it can be. in the beta and all the way through GA, which is not too far away. And I think people will be hearing more about it. Soon, but Timothy, any any first impressions without without giving any details away, without any spoilers, you know. Timothy Ricks: I mean, it's just been really fun to bounce ideas off you and collaborate on like how we can customize how we w wanna work in source, you know? it's been really cool. Matthew: Mm-hmm. All right. Just a just a little teaser there. More to come. Corey Moen: Ha ha ha. Matthew: So Timothy, a lot of I noticed posts from you which I always find really interesting. Like people know you for building Webflow and building Lumos for Astro, but you actually do a lot of projects, like personal projects, you know, for for your family, for your kids and and those around you. And that's I think really interesting to me. So what have you been building lately? Timothy Ricks: Man, all kinds of stuff. I'm just trying to find out like how to use AI in a meaningful way that like creates value Matthew: Mm-hmm. Timothy Ricks: for, like you said, people around me. one thing that's I started working on recently is a app for my kids' school. So right now, the way it works is there's probably like three or four different apps, depending on what grade you're in, that you had to log in. There's one for like the the group chat, there's another for like your curriculum and parents checking in on how their kids are doing. And there's all these different things. so we're trying to consolidate it all into one app. And it's been interesting to think about like the security behind that and Matthew: Yeah. Timothy Ricks: building like social networking features into it and the different roles and permissionings and everything that goes into it. but this is my first time actually building like a true iOS Android app and going through the testing flows and like all the things that go into it. and Matthew: Yeah. Timothy Ricks: it's been it's been really cool. Years and years ago I signed up for Apple developers and and started taking courses to try and learn how to build an iOS app myself. And I just never got far enough. It was but this Matthew: Yeah, same. Mm-hmm. Timothy Ricks: is the first time seeing like something real out Corey Moen: Mm. Timothy Ricks: there and it's been really Matthew: Yeah. Timothy Ricks: cool. Matthew: Yeah, it's interesting. We've got this place where it's finally kind of people who are like us who come from more of a designer, no code background, and it's finally unlocked for us, you know, where we can really deliver these native app experiences. And we don't have to have years of being a subject matter expert in all of the platforms and testing all the devices and best practices. You know, we can just kind of step in. and apply our design and product expertise, you know, and understanding the customers, you know, because we've, you know, use those muscles year after year of doing kind of discovery with clients and customers and then building a solution that fits them. And so this is kind of the sa a similar workflow, just applied into a different Corey Moen: Mm-hmm. Matthew: technology that's now available to us thanks to these new tools. Timothy Ricks: almost like the better you are at just solving high level problems and like making design design Matthew: Mm-hmm. Timothy Ricks: decisions. That's really where the value is these days anyway. So it's really Matthew: Yeah. Timothy Ricks: cool to start using all those skills we built up over the years. Corey Moen: Totally, totally. I think that is like the key thing too that can go missed or people quickly jump to fear when they, you know, spend a ton of time learning a specific tool or specific workflow and that starts to shift in the industry and it feels displacing or something like that. But it's like, yo, it's that h like you just said, that higher level skill, you're just solving problems. You are finding a Matthew: Yeah, yeah. Corey Moen: very efficient way to help clients. You know, or your team communicate to the world through a website or, you know, an app or collaborate or whatever the task is. And this should just now be easier than ever before to do. Right. Like, yeah. And and I think even the time it takes to do it is also interesting how that's shifting. Like I'm I'm curious how you've been tackling that. Like, are you doing you know, sending sending your agents out to work on things, you know, here and there, check in on them occasionally. And do you have like certain times you work on certain things? Like what's your what's your general workflow been like, especially if you have multiple things going at once? Timothy Ricks: It depends on what it is. So like with the app, I'm very new to it and I'm learning it. So I don't feel comfortable letting the agent just run off on its own. But for Corey Moen: Yeah, yeah. Timothy Ricks: something that I've had a lot of experience in, I'll leave an agent running overnight while I sleep like four to six hours and just let it do its thing and wake up to a result in the morning. So it it depends. Corey Moen: So cool. Yeah. And are you are you using certain skills or things like that? Or are you mostly just using models out of the box? Like what is what is the what how is your workflow maybe even change as you've done this more and more? Like you said, some things are newer, some things are more tried and true. yeah. Timothy Ricks: So with skills, it's weird because like I'll write a skill and I'll think it's perfect, and then you try it with a different age or it and or a different context and it all falls apart Matthew: Mm-hmm. Timothy Ricks: so fast. So it's like almost a lot of people have been coming to the realization that actually lessen your skill is better. but I just started getting into hooks and that's been really Corey Moen: Mm-hmm. Timothy Ricks: cool because with a skill or an MD, the agent can like ignore that. But with a hook, you can really start to run scripts that enforce certain things. so something I was working on this morning is can I build a good enough system of hooks that I could spin out hundreds of Webflow components and they're all like spotless? And I I don't have to check the technical side of things. I can just check them from like the design side of things and making sure Corey Moen: Mm. Timothy Ricks: things look good. But functionality-wise and like layout-wise and making sure it works the way Webflow works, like all of that can be written into hooks that are just enforced and I don't have to think about. So hopefully if I get a win in that area I could apply it to like other areas of AI coding too. So we'll Matthew: Mm, mm-hmm. Timothy Ricks: see. Corey Moen: fun. Yeah, I think what's fascinating there, I've had this similar experience of like the skills, or even just it feels like every couple months there's the new hot thing of like it was loops a couple months ago, and then you know, things before that. And like, I don't, I just I think sometimes I'm personally just lazy and I'm like, I just let's just use the model and I'll babysit it when I have to. I don't know if Matthew: Yeah. Corey Moen: I want Timothy Ricks: Yeah. Corey Moen: to. But I think even to that point, and you said it a little bit earlier, it's like. Even if we look back a year ago, there was things we were trying to do that it was like, my goodness, a lot of hand holding, a lot of like, you know, it it just wasn't really executing. And then now I think with some of these latest models, at least in the last few months, it feels like we've turned a bit of a corner on like, now there's more often than not, it's doing things without me even asking, like accessibility and performance and things that it's just like it's at least for websites, right? That I'm like, this is amazing. Matthew: Mm-hmm. Corey Moen: I can just give it the top level thing. Or vent an idea and it's gonna fill in the gaps without necessarily having to pre-build a lot of stuff. but again, I think that it's so hard to say is that applicable to everything, right? Like, or is it just gonna be dependent on whatever the project is you're trying to tackle? yeah, I don't know. Timothy Ricks: Something I've been struggling with lately is getting people to care about the way things are built because now Corey Moen: Mm. Timothy Ricks: that you just see this super polished looking website, you're like, Okay, I'm golden and I don't care about the maintainability because AI is just gonna handle it all anyway. But how d how are you handling that or do you care about like the underlying like way things are structured anymore? Or do you just let the agent run? Corey Moen: Yeah, yeah, yeah, yeah. Yeah. I I mean personally, like Matthew, I feel like we've talked about the same thing quite a bit. It's like I I am definitely somebody that still I think I I live life all often through metaphors of thinking about like how how do we think about these things elsewhere. And like I always I always draw back to like woodworking or something like that. Like woodworkers, you know, use all like it's wild. Go on YouTube and look at like high-end woodworking, all the tools that people have now and like like festool I think is one of the like super high-end brands or something and it's like nuts how much you can do fast, but they still care about the wood. They still care about the process. They still care about like, and I think that shouldn't change with code. Like I still have a passion Matthew: Mm-hmm. Corey Moen: to understand the general approach under the hood. You know, like when we we have even talked about like do we even need tailwind or CSS stuff anymore? Like to just it's Matthew: Mm-hmm. Corey Moen: just all vanilla. It's like To some extent, yeah. Like I st I still think there's something to be said about like, especially in maybe more so the agency and freelance context, like having the confidence to say, I am giving this this solution and I understand it pretty intimately under the hood. but you know, of course the agents are gonna extend on top of that. I think there's the other side too of like balancing though. I I'm more indexing on dependencies. Like I'm kind of personally obsessed with like I want zero dependencies if at all possible in this new age. Like, how can everything just be contained to a repo and that's it? And I don't have to worry about versioning and all these other things. but again, obviously there's boundaries to that, I think naturally. and the other last thing I'd say is like I am genuinely curious about like, will we see a point. Where the models are getting to a point that like we can try to keep those reins of understanding all that it's doing and all that it's writing, but like will it get advanced enough where it it's like, I don't know that it's necessarily gonna invent its own language or something, but it's more just like, is it gonna start doing things in a way that even like staff engineers will have a hard time following? And should we say no to that just because we don't fully understand it? I don't know. That is a very existential question in a way. But I don't know, Matthew, I'm curious, like do you have any thoughts on this? Like code quality and yeah. Matthew: I I do care about code code quality. because okay, first of all, I think that models like I don't know if we've reached it, but there's gotta be a certain point where it kind of plateaus or levels out, like as far Corey Moen: Mm. Matthew: as the advanced. It feels like maybe for web development we're pretty close to that. Corey Moen: Yeah. Matthew: So yeah, we can worry less about having all these super prescriptive skills and and stuff because the models Corey Moen: Hmm, mm. Matthew: kind of just have a lot of that baked in now. so we can worry more about the high level. And one thing I was thinking about is that design is problem solving with empathy and purpose. And Corey Moen: Mm. Matthew: that that's where our value lies in is bringing that to the conversation. And then the execution can be handled by these models that if for for the industry we're in have kind of reached not necessarily a peak, but a place where where it's level and we can like trust a lot of things. But I do feel like in my experience over time, as I've worked on certain projects over months and lots of different sessions, it does feel like it be can become a little bit spaghetti. code just because each agent has kind of done their own little thing or their own little flavor where they're working on one thing and they're only focused on that. and Corey Moen: Mm-hmm. Matthew: so it it becomes our job to like make sure like, no, this is the high level. This is how we're working on this project. This is how it's, you know, the code is organized. This is how it's architected. And I need you to stay in those lanes and we can catch it hopefully in the playing stage before we tell it to, you know, just go and full steam on it. Yeah. Corey Moen: This yeah. It's interesting Matthew: But Corey Moen: too. The way you describe that also is is the thing that for me pops in my head is like, and Tim, I'm curious your take on this, is like we we have a lot of the reason we've emphasized code quality and process order of how we're developing in the past was to prevent, you know, debt, right? Like to prevent a backlog Matthew: Mm-hmm. Corey Moen: or a growing tech debt of bugs and issues that like let's be super real in the pr in the production context we'll never get back to. and especially I remember being like even more kind of maybe over obsessive with this in in Webflow a few years ago of like, hey y'all, like clear your unused CSS and like make sure you're using existing things and like, cause because it's just we're we're not we can't grow that tech debt to a point that we'll just never get to it. However, I think like there's there is something to be said about now we're in this age where like What ironically what popped in my head is I saw some random short on LinkedIn the other day from Linda, Webflow CEO Linda, and she Matthew: Mm. Corey Moen: mentioned the statement of like self-healing websites, which which is stuck with me as Matthew: Okay. Yeah. Corey Moen: just a catchy way to phrase that I think the underlying assumption I made in hearing that was that like maybe the process actually shifts, where as long as things are generally being built, of course, in the stack that you already are working in and things that I think agents already grapple pretty well without much explanation of like Hey, the sites in Astro don't build, you know, a whole Next.js area of it or whatever, like, unless you really have to for some explicit reason. But then get it out as long as it's working, functioning, all that from initial tests. Let's ship it. But then is there some cycle that happens where like every single day you have other agents that are just constantly auditing things for code quality and consistency and design system usage and all of that? And so like. It's just naturally happening, right? Like you'll wake up, you'll ship something, you'll wake up the next morning to maybe some PRs that are now just, hey, I'm gonna clean this up and make sure this is more consistent. You're just approving those. And and you know what I mean? Like, so it's actually like changing the priority order in whence we have to care about the code quality. It doesn't mean it goes away. I think it's still there. It just like maybe doesn't have to be as emphasized in that initial build phase before anything ships. I don't know. Enrant. Timothy Ricks: I could see us getting there, but I don't think we're there yet. Just from personal experience, this past week, we tackled or I tackled like a migration of sorts. And Corey Moen: Mm. Timothy Ricks: it we the first version was just like the spaghetti code version. Like nothing was components, everything. And then we try to do the refactor after and it just didn't. It wasn't as clean as just going straight to the the right way the first time. So it was much, Corey Moen: Mm, interesting. Timothy Ricks: much harder to get there. It took many more steps. So Corey Moen: Mm. Timothy Ricks: maybe we do get to that point one day, but like just personally, I I haven't seen it work that way yet. Corey Moen: Totally, totally. Matthew: I I agree. I feel like people right now are living trying to live in the future where they're like, we'll be able to refactor things easily in, you know, in six months. You know, the models will get there. You know, it'll be easy. So we don't have to worry about it now because it'll be taken care of. But it's like, Corey Moen: Right, right, right. Matthew: even when even if in six months that is true, it's Corey Moen: Mm. Matthew: gonna be potentially in a big refactor like that, there's a lot of risk. And so it's like you can manage Corey Moen: Hmm. yes, for sure. Matthew: that risk of, you know, bugs and breaking your product, especially if you have at like actual customers and it's not just like something internal. You can avoid that by, you know, building the best quality you can now. Corey Moen: Totally. Totally. Matthew: there's there was a live stream from Matt Pocock and with potato And it was really good. And speaking of analogies, her thing is that, you know, it's not a software factory. It's that it you're a Michelin chef, you know, Corey Moen: Mm. Matthew: and and you know, speaking of like code quality, it's all about like as food's coming down the line, like there's there's a check at the you know, at the end, you know, there's a line check, you know, and then the sh chef is going to look at everything before it goes out. Corey Moen: Mm. Matthew: And There's this concept of having agents that are Corey Moen: On that. Matthew: reinforcers, you know, who basically are doing that like quality check of all the other agents' work before anything goes out. You know, so Corey Moen: Mm-hmm. Matthew: having those kind of gates and then how can you scale things up? it was really interesting to hear the analogy kind of sprinkled throughout the conversation because because it actually worked, you know, from like, you know, every chef like just like Us, you know, it's like if you move from job to job, you're going to take your your AI, you know, skills and tools with you wherever you go, just like a chef, you know, has their own special utensils to, you know, you're you're trying to take what you do in in one restaurant and duplicate it and like franchise and build out, but maintain that kind of Michelin star quality wherever you go. so a lot of good principles in that talk, I suggest everyone check that out. Corey Moen: Super cool. Yeah, it's it's funny too. It just even I love the description too of like I think we especially also feeling like in Webflow, when you do build with frameworks and things like that, it does get to a point where there's that certainty of just like you build a thing and you don't even almost need to check responsiveness. You just know it's gonna work because you're building that that stuff's already been figured out. You know what I mean? So like still starting with that foundation totally makes sense. yeah, anyways. Matthew: Okay. Corey Moen: Sweet. I mean Tim, would you be open to sharing maybe sharing your screen and showing some of the some of the projects and stuff you've been been working on? Timothy Ricks: Yeah, sure. let's see. So this is a game that I'm kind of letting my kids direct. So my daughter drew this kind of character and asked Claude to turn it into a 3D model. And it wasn't connected to Blender or anything like that, but it drew her it recreated her drawing really well as a 3D character. And Matthew: Mm-hmm. Timothy Ricks: I've just kind of been letting them, you know, do whatever they want to do with it. So it's grown into a hodgepodge of many different games. So like you stop these bad guys from eating. The gingerbread people. And then you also can kind of like build stuff. Corey Moen: Snowballs. Timothy Ricks: yeah, yeah. You can build stuff, and there's little details like if I build a fire here, I can like roast a marshmallow, and the longer I stay there, the darker it gets and all that. Matthew: Uh-huh. Timothy Ricks: but yeah, it's there's pretty much a diner dash recreation in here. there is this little toy workshop. Where you build toys and ship them off in here. there is over here. You can at the edge of the mountains, there's like you wouldn't be able to go that far. but my daughter was like, I actually want you to be able to go that. So now you can walk as far as you want in any direction, and it just keeps generating more map. this is the little church, and they're just used to getting a candy. after church. So they said you gotta get a candy. So now you get the candy and eat it as soon as you leave. Matthew: Mm-hmm. Timothy Ricks: you couldn't go in any of these. So now you can, you know. just lots of little it's been fun to just watch their imagination, whatever they want to do, you it can pretty much just do and just watch it evolve into whatever they want. Matthew: That's cool. how how did this start? Like how what was your process w working with your your kids on this? Timothy Ricks: Well, so she sh send that drawing. She made that drawing and she said, Matthew: huh. Timothy Ricks: Hey, can you make this into a game? So I just took a picture of it and then gave that to Claude and said, Make this a game. And the first past was pretty much this. Like all the styling, I didn't have to do any change on the styling of things. So I don't know where it pulled like these style references, but the style was great. And the these houses were there. And then V2 was walking into the houses, building this little kitchen. building the other Matthew: Mm-hmm. Timothy Ricks: stuff and we've just been slowly adding on to it from there. But I've just run Matthew: So she has feature requests for you and then you work on it together? Corey Moen: my Timothy Ricks: yeah, yeah. you can build like stuff on here. I I think you can build like a slide and go down it. yeah, there we go. Corey Moen: Like was that something you asked for? Like like being able to build? Timothy Ricks: Well, what's funny is she she said, I want a tree house and I want it to have a slide and I just said the tree house and before I could even say and add a slide it had already added the slide. Corey Moen: Yo. Creepy. Matthew: Mm. Timothy Ricks: Yeah. Yeah, I know. yeah, it's it's been fun. She she has feature requests and stuff and yeah, Corey Moen: First person, my goodness. Timothy Ricks: yeah. I did one for my father in law. it's a hunting camp game. Corey Moen: Mm-hmm. Timothy Ricks: And essentially like it turned him into a three D character and I it I took a picture of his boat and it recreated his boat in three D. So he can walk in the boat, he can drive the boat all around to like the camp which is on the water, go inside, decorate the camp, go outside, catch crawfish, shoot deer, cook, go to the woodwork shop, build a bunch of stuff, put it wherever, like just completely open world game. Yeah. Yeah. Corey Moen: Unreal. And is this like was this were were these the overnight scenario? Like, you know, it it ran for hours to get to this point. yeah. Timothy Ricks: No, like so the Hunt and Cant game, I used the ax the Axolotl game as a starter and just said reskin it. And so that Corey Moen: Mm. Mm. Timothy Ricks: came together in like maybe two hours max. this axolotto game, the first version of it, came together in less than an hour, I think. Cause I just ran it on astra fast mode and I would have multiple things going at the same time, like multiple chats. So 'cause she's not gonna sit there and wait for hours as this works. Matthew: Mm-hmm. Corey Moen: Yeah. No, no, no, no. Yeah. Timothy Ricks: So I'm just like, yeah, it it came together really, really fast. So Corey Moen: Unreal. Timothy Ricks: That's cool. Matthew: Does she understand anything about like AI and like how this is happening? Like it's just like I talk to the computer and it just does builds what I want, you know. Timothy Ricks: I don't think so. Yeah, sh yeah, she just thinks the computer's magical, I guess. Yeah. Matthew: I probably most of us have an a point where it's like computers felt magical, Timothy Ricks: Yeah. Matthew: you know, until we understand that it was something that was designed and built by other people. but yeah, it's interesting to see how the younger generation is just gonna grow up with this just being normal, you know, it's just hey, speak and then there it is. Timothy Ricks: Yeah. Yep. It's right now I see it inspiring her art. Like she wants to right now she wants to make a little clip, like a little short video. so she I was like, well, you're gonna need characters. So she drew like 10 different characters and we like made them through D and all that stuff. So right now it's I see it as like helping her get to the point that she can't get to right now and hopefully like she'll like fill in some of those extra skills as she like gets interest in different areas. because I remember as a kid, like it was I loved making videos, but it was all very manual. Like I would draw out every little frame in Microsoft Paint and then I'd stitch them together in movies and like slowly you get better tools. But like it's cool to like just see your vision a little sooner. And I think that inspires you to like want to keep going and keep wanting to learn more. Yeah, Matthew: Mm-hmm. It's a faster loop, right? Timothy Ricks: yeah. Matthew: Faster feedback loop. Which is great for any kind of creative process. The the faster you can make that creative loop and then have the friction elsewhere. Yeah. Timothy Ricks: was just gonna say as soon as you hit a wall that of something you can't do with AI, it like makes you want to actually dive deep and understand how it works and yeah, get more custom. Corey Moen: Totally. And it's I think it's it's wild too that it's to to the point I before I I there's a lot of things that would come to mind and it's like not even worth the energy to try to see if it's possible. Whereas now it gets to a point where like, well I could at least just float start a new session, float the idea, leave. I'll come back in a day or two and see what it thought about it when I remember. that I that is pretty fascinating, you know. it it's I think the irony is. The more people I talk to that are maybe not as intimately in the space that we are of like, you know, building websites and all that, where I think this naturally lets us leans us on the bleeding edge of a lot of this is like there's just so many assumptions of like, why would I even ask that? Like AI can't do that. And it's like I you would be surprised. just Matthew: Mm-hmm. Corey Moen: get in the habit of asking. You know what I mean? Because I think sometimes, yeah, even if it can't do it. off the cuff, you know, whatever the thing is, like it might have a lot of ideas on how to start elsewhere, you know, and at least is like a sounding board for you in that way. I don't know. Yeah, super interesting. Matthew: Mm-hmm. Timothy Ricks: Yeah, I definitely see the limiting beliefs, even from web designers who are like, I just don't think AI can build this section. Like and it's like we're we're past that point at this part, Corey Moen: Totally. Yeah. Timothy Ricks: but yeah. Corey Moen: It will find a way. Whether Timothy Ricks: Yeah. Corey Moen: it built it in the way that is ideal, who knows, right? Timothy Ricks: Yeah. Corey Moen: but yes, yeah, yeah, yeah, Yeah. I even even like the way it can fill that gap. I think something else I was like going down the rabbit hole in the last few months just in personal time was like trying to like think of other ways to net that gap between I have content and I just wanted to map to a system of possible layouts and every and all that. And then like I feel like I got to a point even recently of like, I don't know, the models are getting faster, better and cheaper, like very quickly. And so is it is it even worth the time of trying to have some other system that fills that gap or you know, whereas like a year ago or even six months ago, I was like, I I don't I don't think it's worth spending the cost and time of an agent to fill the gap. But now I don't know. Now in these recent few weeks or months, I'm like, maybe maybe it is worth it, right? Because it can deal with so much of that gray nuance that is hard to describe or pre-plan. depending on the situation, especially with like marketing sites and some of that. so I don't know. Yeah, it's it's interesting. And I think like all that to say too, even thinking about, you know, team members if you're in-house Matthew: Yeah. Corey Moen: or clients if you're a freelance agency is like we also have always indexed on the like pre-building of things that make it easier for them to use and manage the site. And like I don't I it I still don't know if I'm a full believer, but like I definitely used to doubt that's like, well, yes, it's of course so easy if they can just ask an agent to do whatever it is they want to do, but there's all this other unpredictability in that, and that amount of unpredictability is shrinking fast as well. So I'm like, then I am like, okay, maybe it is just easier for them to ask, just like they would ask me. Cause like how how often have we also had that feeling of you built spend all this time building this beautiful system. So that they can do it and then they don't do it. They still ask. You know, like they won't watch the video, they won't read the docs, like and I you know, we're all busy, life is busy. and I it's I think the deep irony in it is even if even if it is as easy for them to ask the agent versus you, they'll probably still ask you. But maybe maybe that's just what they see it as, and it's actually the agent under the hood. I don't know. You know, like this, that's what I mean. Yes. Matthew: You just need to make a Corey bot that looks and talks just like you and they can think they're asking you. Corey Moen: Yeah, yeah, yeah. Or like or like more recently, and and I'm thinking about all this and what you said, Timothy, of your daughter just being like, I drew a picture, make me a game. And it was almost like you you could have just had record on in the app and she was just saying that. You didn't even have to retype it. And it just went straight through to it. You know what I mean? And so like I've had that same feeling of like on calls with the team or or you know, whatever, and like Sometimes I've like, can I just record and just whatever you're asking for, whatever the feedback, just like straight pipe through. I don't need to be the the intermediary always. But to be clear, I think sometimes it's very still necessary because we have to interpret all the soft context around it. But yeah, it's it's interesting how that gap is just changing pretty quickly. between like, yeah, feedback and ideas to execution. the other thing that comes to mind here is that is the instantaneous part of it is what is maybe the magical part of it. And so, like in in the design development realm, I have wondered about like, especially if y'all saw, you know, when Jev that new model came out and it's like super hyper fast, it's deterministic, all of this. Like Matthew: Are you asking me if I remember two weeks ago? I mean, no, I don't, but okay, continue. Timothy Ricks: Yeah. Corey Moen: I mean fair, fair, fair. I don't. is is, you know, remember even a couple years ago and before when the early stages of coding was autocomplete, right? Like as I'm typing, then it's just Matthew: Mm-hmm. Corey Moen: filling out strings. Like we're still I s I am I am waiting. It feels inevitable, we'll eventually see this with the with design, right? Like you are designing things and you don't even have to ask. It's just like as you're moving shapes, it is starting to understand and then it's going to J you know, type ahead your design. And in the same way, this feels kind of like that essence, if that makes sense. It's like I'm gonna capture the general thing, I'm gonna wire some boxes, and it's just gonna fill it in for me. I don't even need to ask or prompt. There is intuition that happens from the models in that. I I again, I think a lot of people are going after this. Maybe this is even a segue of like Timothy, you mentioned subframe, I think it was. a tool, you know, and there's like I know last week Matthew we were talking about paper dot design became pen dot design. I'm still fascinated Matthew: Pin. Pin.dev. Corey Moen: to know why the name change seems like so so random. But you know, there's a lot of different tools coming out of the woodwork trying to generally do this, like just design on a canvas and it is code, but also it's gonna help you fill it out, right? Even Framer with their agent, I think, has been really interesting on the tools of like. you design desktop and you say make it responsive and it just you know, fills it out like all of that in the isn't that is in that same bucket. So I don't know. Yeah, Timothy, have you have you have you played with some of these tools like paper or even subframe you mentioned or any any thoughts on this general idea of like type ahead design? Matthew: Mm-hmm. Timothy Ricks: Yeah. it's the it's the thing to solve right now because I haven't really touched paper before. but I'm really interested in that because more and more of my design is actually happening in code. And I was never like a super strong designer to begin with, but Corey Moen: Mm. Timothy Ricks: it's just so helpful to have something to react to and then I can evolve it from there. But it's much harder to play in code than it is in Figma, at least for me anyway. Like if I would just want to freely drag things around and like experiment. you don't get that instant feedback. You're sitting there waiting for the agent to come back with the next thing. So Corey Moen: Totally. Mm-hmm. Timothy Ricks: there's a a browser extension, a Chrome extension now that you can actually it enables like Figma on your website. So you can drag things around, like experiment, and then see if it that's what you're wanting, and then screenshot it, give it back to the agent. and then it is let me see. Corey Moen: Cool. What's the extension? Matthew: Is it is it squints? Timothy Ricks: Let's see. Sorry. I don't think it is. Corey Moen: That's a great name if that's what it is, Squints. I love this. Matthew: Squints is one that we shared on No Code Supply Co. the other day. It's a Corey, you may not even know that. Yeah. It was Derek and I posted it, but Squints Corey Moen: weird. Yeah, yeah, you know, it just the machine. Wait, squints plural or squint? Matthew: dot app, yeah. Corey Moen: there it is, okay. Interesting. Matthew: Yeah. So this is a Chrome extension that adds like a UI in the browser. Yeah. Corey Moen: Do you wanna share, Matthew? If you have it up. Matthew: I don't but I don't know that this is what you were talking about, Tim. Corey Moen: Yeah, yeah. I mean it I think again there's there's I've seen many instances of this general concept too. Even in in cursor, right? There you can like quick modify things or you know, claude design. There's also like you can mod things and c recolor and reorder. it is interesting, you know, like it feels like you said, Timothy, like a lot of these things are just inevitably You know, a lot of people are trying to solve this. This is one of the hot things at the moment. that I have yet to see like the perfect thing that balance all balances everything. But like, yeah, this one looks interesting. Matthew: Mm-hmm. So it's the idea here is you're bringing Figma, Figma like tools directly into the to the browser, into Chrome, you know, so you can Corey Moen: Mm, mm. Matthew: make your adjustments, play around, right? And then you can capture those changes and then, you Corey Moen: Mm. Matthew: know, use your agent to actually go and make them. Timothy Ricks: If you look up design mode dot app, that's the other one and it looks very similar. s solving a very similar thing. Matthew: Yeah, I think I've seen this one. Yeah. This yeah. This one yeah brings more like a full design panel to the side. Yeah. Corey Moen: yeah, this is like Webflow style panel in an extension. Interesting. Matthew: Yeah. Yeah. Yep. And then you can, yeah, see there right there you can copy those as a prompt or send directly to your agent from there. Corey Moen: Yeah. Timothy, you mentioned like Canvas and stuff. I don't I've never actually tried it, but I remember Figma at config announced I can't remember what they called it, but so you supposedly can render your actual project like localhost in a frame. or was it directly from GitHub? I forget. All the the detail I remember is that they you could render it directly on the canvas, and then when you option click to drag and duplicate, it's it's a new branch. so Matthew: Mm-hmm. Corey Moen: like it was trying to replicate this exact same experience of like the classic designers in Figmas, like dupe, change, dupe, change, dupe. And you have this like, you know, downward stream you see on X or whatever all the time. but they're all it's all real code. It's all actual branches. And so it's just another way to to express, you know, that same workflow, but like you said, in a way that you're not waiting for things, you're just instantly getting feedback and adjusting things along the way. I don't know. Matthew: That tighter feedback loop. Yep. Corey Moen: Yeah, yeah, yeah. I think what's interesting too is like with some of those tools, I don't know if they quite work this way or not, is what we were talking about earlier, code quality is like I don't I also don't necessarily want, whether it's me or honestly even a a teammate, a teammate or a client, like I don't necessarily want to just freely do any pixel value or anything. Like ideally the things I'm being shown that I can adjust visually and move around is still innate to my design system, right? Like only give me specific values. Matthew: If you're if you're dragging that slider, it's snapping to what your system values are, right? Corey Moen: Bingo, this, right, right, right. And it should feel fluid like it's one off, but like or like if I'm seeing it if I click on some text, I don't want change the size, I shouldn't just get whatever. You know, it should be like here's your six heading sizes or whatever. Matthew: You're not gonna do thirteen point five pixels? Corey Moen: What's that? Matthew: You're not gonna do thirteen point five pixels for your subheadings. Corey Moen: I mean, yes, if it's equivalent in REM. Yeah, I don't know. It's super interesting. I the other thing I'm curious on y'all's take is like what we just mentioned is localhost, right? So like even if and we don't have to get on the source thing specifically, but I also putting the like the teammate hat on, giving other people than myself as a developer the control to manage the site. Local host is an intimidating thing, in my opinion. It is confusing. It is like it is so innately tied to the way that Git works, that you know, you pull things down and push things up and you have branches. And I just think obviously source is one of several tools abstracting some of that, right? Like there is a publish button that is, you know, from my understanding or at least guess from seeing the visuals, because I have not used source, is it's just abstracting the idea of merge, right? Or or or or whatever is happening on top of Git. But You to work locally, you would really still need to be on a computer. What about like you know, somebody that wants to do something on their phone and just be like, again, ask the maybe it's just asking the agent or whatever, like, hey, go update this thing, whatever. I think there's still that gap that I'm not truly understanding with like source or some of these other tools is like if I'm handing this off to a client or to a team because I'm now going somewhere else or whatever, like Yeah, how do y'all think about that? Like how how do the how do the the non technical people fit in these newer tool workflows in that way? Timothy Ricks: So I can share because I I've done this with a client a couple of weeks ago actually and had a lot of success. But the piece that was missing was the collaboration piece. So I think that would bring Corey Moen: Hmm. Timothy Ricks: a lot more complexity into it. But basically the client had built the website for them and they get requests every week from all their different campuses. And so for them, having to figure out how to do that in a visual builder was just too much. And there was things that Like one of the edits, for instance, was add social media links to the staff cards, but only for the lead staff. And so if you break that request down, you have to add the CMS field, you have to find the SVG social icons, you have to add it into the card, you got to link conditional visibility so it only shows up for certain staff. There's like a lot of things. And even when I trained Corey Moen: Yep. Yep. Timothy Ricks: this person really well how to do web flow, it was just a a big blocker. And all that can be one prompt. so Corey Moen: Bingo. Matthew: Because they're not in Weblo every Timothy Ricks: moving yeah. Matthew: day, right? Like how often do Timothy Ricks: Yeah, exactly. Matthew: they touch Weblo? So trying to retain that what you're teaching them is difficult, right? They don't live it every day. Timothy Ricks: Exactly. Yep. So basically now they're copying the requests that come in almost word for word what they're getting, throwing it into AI and then reviewing the result. And local host for them is basically like staging. So Corey Moen: Okay. Timothy Ricks: once it looks good, they're like, all right, push this up and then it's done. So we looked at a couple keywords like push and pull. one thought process that was really helpful for them because they're familiar with after effects and video editing and all that stuff, and they're Corey Moen: Mm. Timothy Ricks: trying to they really want us to understand it. They're like, what's the difference between like GitHub and Cloudflare? And like, how does this fit together? And like, so Cloudflare is kind of like where your exported video goes. And then GitHub Corey Moen: Yeah. Timothy Ricks: is like what holds your project files. So you can save it all in Cloudflare, but then if you need to make a change later, you won't have the editable file. It'll only live on your computer. So you also need to store it somewhere else. They're like, that clicks and like bingo. So it's just finding ways to explain complex problems to people in a way that they'll understand But they they definitely see the value out of it. Like you could tell their eyes light up when they realized how much easier it was gonna be for them to make those edits going forward. Corey Moen: To just ask. Yeah, Matthew: Mm-hmm. Corey Moen: yeah. So cool. I also, quick aside shout out, Cloudflare. I only very recently realized like I had moved all my domains, personal domains there, I don't know, a couple years ago when Google Domains went down. Yeah, yeah, Matthew: you you're using them as the registrar. Okay. Yeah. Corey Moen: yeah, for sure. When Google Domains went away, which I loved because it just mainly made setting up Gmail so easy, custom Gmail. Matthew: I've been on hover for like, I don't know, decades now. Yeah. Corey Moen: man. And so I I migrated a personal site recently and I and I realized that almost all of Cloudflare's dashboard is responsive. So I can again, like I've kind of been maybe too obsessed in my I Matthew: On your phone. Mm-hmm. Corey Moen: don't know if it's laziness. I make the excuse that it is I don't want to be on a computer anymore in the evenings. I just want to relax. And if you know it I if I'm gonna check my phone. Yeah, Matthew: Are you gonna be pre ordering that iPhone duo so you can Corey Moen: that's true. Ooh, I didn't even think about that for like editing. Matthew: Flip it open, do some work, close it. Corey Moen: Yeah, long story short is like I did I did this migration. I only touched my laptop once because I could do the DNS changes and the workers and all that just like from my phone. And I was like, this is nuts. Anyways, so yeah, shout out Cloud Flair. Also just very generous free tiers and things like that. Matthew: If Anthropic bought Cloudflare, would it be Cloudflare? Corey Moen: That would be ultra cheesy. Matthew: You you you can propose that to the team and you know, send me a check. Corey Moen: right am I with webflow buying tools and adding flow at the end of everything, right? an aside, I think something I want to point out that you called out there is the collaboration part. Is that like is is inevitable. This is also something that's been Matthew: Mm-hmm. Corey Moen: then on on the mind recently is how ironic. You know, we've overbuilt a lot of tooling over the years in across development for humans and all of this. And then it feels like there's certain things that are boiling back up that have always been around, but maybe they just have a a fresh, simple take. And one of them is commenting tools. Is like, you know, we used to use when I was on the team at Webflow, we used markup to IO for a long time. Then we kind of shifted to commenting directly in Webflow. Long story short, like being able to just see the site. Leave comments for yourself and others. Every all feedback is captured in one place. There's many different people, even in the webful community, that built their own tools doing this. And I'm like currently very fascinated by like, what if could that just be the simplest essence of managing a site? Is like you're just go opening the site and just like you leave a comment on something you want to happen for somebody else or for an agent to go start doing. Matthew: yeah. Corey Moen: And then like that. is the start and stop of most of what you would need to do. And even in what you described, Timothy, is like the the the nuance of of the social icons, right? Only for one case and the wiring up the CMS and all of that, that's again, can the agent just fill that gap? Can it fill that gray area? I do still think it is really easy to say those things at a high level. And then I feel like the more I've started to go down that hole, the more I'm like, but there's still all these little like governancy type things that still matter. You know, not anybody should be able to come in and just be like, make the homepage pink, you know, and then no one else is like gating that or Matthew: Denied. Corey Moen: making, you know, there is absolutely still you need to have some like control about how this goes about. But I'm just always fascinated and maybe too obsessive on simplicity and like I just wanna, I I don't want the tools. I just want like what is the most simple way and also like closest to the actual site itself, because there's this whole other side of like, hey client. Or hey team member like that maybe hasn't used an AI and all that, like I don't necessarily want to ask them to like, all right, go sign up for Claude. Now this is how you use the Claude Code des or app. And here's what a repo is, and here's what a you know, again, like you said, yeah, Matthew: Connect this, install this skill. Yeah. Corey Moen: yeah, yeah. Like I I I in in some sense, I just want that abstracted for them, at least to a degree. So anyways, and Rant, I just think it's fascinating what you said that the collaboration aspect is still so important. In almost any use case of website, even your personal portfolio, I could argue that like you still would want to collaborate with an agent in this sense at the very least, but you might also have cases where you know you want a friend's feedback on like, what do you think about this case study before I ship this or something? Right. Like we are humans and I think we connect so innately through that, even if it's async through a commenting system or Slack or whatever, I think that still matters. We care about. the other people in our lives and their opinion, whether it's in a team or a family or a friend context. And so like finding ways for that not to get lost in this kind of like shift, I think is so valuable. I don't know. Hot take. Matthew: Not a hot take because connection is what it gives us energy. I mean, it's Corey Moen: Mm. Matthew: you know, there's a lot of mental mental and health benefits from connecting with other people. So Corey Moen: Mm. Matthew: we're we're going to retain that. You know, we might find friction with these AI tools for a time, but like you said, it's it's kind of inevitable. This collaboration layer on top of all of this. It's Corey Moen: Mm-hmm. Matthew: just kind of how long does it take to get there and for us. to kind of figure out what what does this actually look like, what actually works in practice Corey Moen: Mm-hmm. Matthew: with people who are both technical and non-technical, you know, all in the same place. Corey Moen: Totally, totally. Love it. Matthew: Tim, any any parting thoughts from you? Timothy Ricks: No, this has been great. It's been so great catching up with you both. Thank you. Corey Moen: Yeah, Matthew: Yeah. Corey Moen: likewise. Thanks for coming. Thanks for showing some of what you've been jamming on. Matthew: Great chat. Yeah. Timothy Ricks: Yeah, of course. Glad to. Matthew: All right. Well, thank you everyone for joining us. a couple of new things. We just want to say if you go to No Code Supply Co slash feedback, you can send us your thoughts. What did you think about the show? Do you have anything you would like to add to what we said? Any recommendations, nominations for guests or topics, we're all ears. And then lastly, you can now subscribe to No Code Supply Co. So just go to the site slash subscribe and click that button. And you'll get to hear whenever there's new stuff from us. So all right. Until next time. Corey Moen: Thank you. Matthew: See y'all. Corey Moen: Bye.