Taylor MacDonald: All right. Hey, welcome to Blood, Sweat and Tokens, the podcast where we talk about AI tooling by putting it through the paces in real product development lifecycle. We're building real solutions to real problems out in the world, but we're using a lot of AI tooling and other agentic strategies to get that work done. Sean, how you doing today? Sean C Davis: I'm doing good. I know we've been we've been talking about this show for a while, so I'm excited to dive in and see what we can bring to life. Taylor MacDonald: Yeah, I am too. I guess we should probably introduce ourselves since this is the initial episode. My name is Tabor MacDonald. I'm a partner and CTO at Ample. We're a digital services consultancy ⁓ and marketing agency in downtown Cincinnati. We've been in operation for about 20 years and every day I wake up there's something new to learn. So, what about you? Sean C Davis: ⁓ that's a good idea. I'm Sean C. Davis and Taylor, well we met, we met in Cincinnati and I started working for Ample, what, I don't know, maybe almost 10 years ago now or something like that. It's been a while, it's been a while. But I, ⁓ yeah, go ahead. Taylor MacDonald: ⁓ Can I tell you, my son turned 13 today. That's how fast time goes. Sorry for the interruption. Sean C Davis: Holy smokes. That's going. Yeah. And so since we're going to Ample, bounce around a little bit. today, I am at Netlify in developer experience. yeah, and we've pivoted into the AI space and trying to build our platform to be the easiest and ⁓ yeah, easiest platform to ⁓ build and manage applications, particularly for folks who want to use agents. And so yeah, really, really excited to get our hands dirty and solve some real problems. Taylor MacDonald: It's exciting. Have you guys been hacked yet? Has anybody infiltrated your Google workspace yet? Sean C Davis: Let's, ⁓ my desk is made of wood, right? So we'll knock it. I have, I have, but it hasn't gotten into the company. But we'll save that story for another time. Taylor MacDonald: So. ⁓ my goodness, I can't wait to hear that story. That sounds terrifying. ⁓ Okay. Well, listen, so when we first started talking about this, Sean, ⁓ you know, you and I have talked shop for all these years. One of the common things that we have implemented. at every project throughout the course of our career, at least mine, you maybe until you moved into the ⁓ more of the solutions engineering world, is CMS's, content management. Everybody hates their content management system. Would you agree with that? Sean C Davis: Yeah, that's accurate. ⁓ I mean, I've built, I think, maybe five CMSs from the ground up. You've probably done something similar too, right? Taylor MacDonald: Only one and I mean I felt like a sadist doing the one to be honest with you. I can't imagine doing five my lord that's it. Sean C Davis: Hahaha I mean, it's five iterations on the, you know, tearing it down and building it back up, but always just like, well, the things in the wild are my clients hate them. So let's solve this specifically for our clients. And I don't know. It's been a, so yes, it's been a huge pain point for almost anybody I've worked with over the last 15 years. Taylor MacDonald: Yeah. Let's talk about why they suck so much. Why do people hate their CMS? What would you say is the number one or number two, you know, couple of really prevailing frustrations that stick out to you as you reflect on all those years of wrangling CMSs? Sean C Davis: Yeah, I gave this conference talk in I think 2020 or 2021, and I feel like it still holds true. the theory that I had was that every client is coming with a unique perspective and has a different set of needs that they require for each. site or project, but also for each page. And I look at it as this spectrum of like, Hey, if this is super rigid, then having a form to fill out the fields is fine. But then I don't have the flexibility. So there's like, there's the speed and there's the flexibility and that spectrum, you know, where, where the the sweet spot is, is very dependent on, I mean, it has a number of different ⁓ dimensions to it in terms of at the page level, at the project level, at the individual editor level and their comfort and ability and creativity or comfort around creativity. And I think that to build a product, at least, in the past with the tools that we've had in the past that you have to, I mean, a product has to bring some level of opinion in terms of where are they going to fit in that spectrum. And so they might, you know, you bring a, you know, to name names like, ⁓ WordPress was really popular because it was really simple but could be extended. And then you had those headless CMSs which were like, it's super flexible, you do whatever you want. And then it's up to you and me on the front end to decide how these things should work themselves out. And it's just, there's never really been a silver bullet. And so what's good for one page or one person or one project is gonna fail for. some other scenario. Taylor MacDonald: Yeah. think that you, okay, so you've really zeroed in on a major challenge that we have. So as a digital services agency, one of the biggest problems we have when we get in front of a client and we're going to... pitch them on some sort of solution, which inevitably involves some sort of content management or other platform as a service integration, is how technically savvy that person is. Who is the team? And often, we're talking to stakeholders, right? So we don't know the marketing person that is tasked with managing all this stuff. And depending upon their level of competency, their comfort level with things like HTML, CSS, or how about their... ⁓ Sean C Davis: Mm-hmm. Taylor MacDonald: how about their ⁓ kind of ambition? We've got a client right now, he's using Figma Make to generate React components and literally copying the entire code and pasting it into a text area. It technically works, but you can imagine how much heartburn that gives us because it's like, you know, performance problems, proliferation of different script tags. I mean, the ability to totally jack up your front end is, I ⁓ mean, it's here, it's terrifying. ⁓ Sean C Davis: Yep. Taylor MacDonald: I don't blame them at all. know, like one of the things that always kind of blew my mind is we think about these things in terms of like page types, but that's not how a lot of these companies need to operate. Like a lot of companies, or let me say this, know, 15 years ago we would build CMSs around the concept of different page types. You have a homepage, you've got a product page, you've got a, you know, testimonial page, maybe a contact page, that kind of thing. Sean C Davis: Mm-hmm. Taylor MacDonald: That's all well and good, right, until the client needs evolve past those very discreet use cases. So they're like, hey, I'm gonna roll out this new landing page for this campaign I got coming up this weekend, but I don't really have a page template for that. So I'm just gonna repurpose the homepage template. Now I have two homepages. Not to mention it's confusing from like a ⁓ terminology standpoint. You look at that and you're like. Sean C Davis: Yep. Taylor MacDonald: There's two homepages, which is the homepage. You ever run into that kind of scenario and how do you sort of talk about that? I know you're doing a lot of, you know, consultation and solutions engineering over the last couple of years. And in fact, before you came to Netlify, Humblebrag, you worked for, was it Stackbit, which is aiming, well, it was at least at that time, aiming to solve a lot of the technical nuance and reduce the barrier to entry through visual editing and things like that on that platform. Sean C Davis: Yep. Yep. Taylor MacDonald: But yeah, how do you think about that? And like if you were going to build a CMS today, what, I don't know, what revelations do you have, you know, along those lines? Sean C Davis: Yeah, because I think the stack ⁓ bit was kind of coming in. The opinion we brought was that the editors want to work more visually. They want that... It was more like I want the Squarespace experience when I'm the user or the editor, but I... But on the developer side, we want full control so that as we understand where our users fit in that spectrum, we can design the system to serve them. But I think, back to your point, it's like generally we've found that those page types, those rigid page types are too limiting. And then on the other hand, There you give somebody full control to drop HTML and CSS into a Blank text area your body of your page and you've got this other problem. And so I think I Mean we could dive into like some some of those problems that we've seen with ⁓ You know even in the last couple of years, but I think what I'm seeing lately is spending The place I'm seeing people ⁓ be more successful, and kind what we're starting to do at Netlify, just on our marketing site itself, is focusing the energy on the system, the baseline tooling, and then letting your stakeholders build on top of that. And so that's what I've been talking with our teams a little bit more about is, okay, rather than, what's good example? Okay, so yeah, I've got this little microsite or landing page that I've got to spin up and I want it to look a little bit different and unique than everything else on my site. So how do I do that? I'm gonna bring in the engineering team and the design team to make that happen. And... What we've been starting to shift to is rather than say, okay, cool, yeah, we'll build that for you. It's gonna take two to three weeks. We say, you know what, instead of taking that time, let's make sure that the rules of the system are built in a way that empowers this team to be able to do that so they don't totally screw up the design and make it look terrible, but they also have the power to make it flexible. And I think ultimately what that leans on is, them using more like natural language and some sort of AI agent that is then informed by the system itself. And so the engineering becomes more about building the guard rails and the design engineering as well about those guard rails. Taylor MacDonald: Yeah. Yeah, I love that. ⁓ You're speaking more to like the primitives that are available within the Netlify ecosystem and stitching those together through some sort of orchestration layer, right? Is that a fair assessment when you talk about the core building blocks or are talking more about like the actual design system that would be implemented through, you know, the front end engineering? Sean C Davis: Yeah, I think it's, I think it's a combination and really regardless of platform or tooling in that. your, your, your lowest level is the, it's the code itself, right? Like the individual components and how they work and function. And that's the extent to which your, your site can operate. And then, so I kind of look at that as like, that's your foundation. And then you've got this layer on top of that. Well, if you jump all the way to the application of it, it's your marketing team or whatever working with some sort of agent to compose those components into a page. So there's gotta be this layer in between there which says, okay agent, are allowed to do this, you're not allowed to do this, so that we're, exactly, exactly, exactly. Taylor MacDonald: Right. Yeah. here are the constraints, yeah, so you can kind of keep the user in the box, but they still have the creative control and flexibility. Okay, so you touched on something that I've been really passionate about these last six months or a year or so, we've really been leaning into the AI tooling, which is this concept of friction. If we start the discussion here around the idea that everybody hates their CMS, well, why? Why do you hate your CMS? You mentioned visual editing in the case of Stackbit, which that is a Sean C Davis: Mm-hmm. Mm-hmm. Mm-hmm. Taylor MacDonald: I think we've gotten to this world whether in, you know, Miro style infinite canvas applications or we're talking about Figma or Webflow or Squarespace where you literally have more, it's like a design centric toolkit that lends itself well to the pipeline, you know, ⁓ naturally becomes ⁓ code adjacent or through tools like Figma, MCP, things like that. We can actually extract those components out and generate actual realistic production level code, which is new, right? That's awesome. That's amazing. But I want to get back to the thing, why does somebody hate their CMS? My argument is, well, beyond the fact that it just completely sucks to type in a little tiny box, why do those boxes exist? They exist for structured purposes, right? So if we're programmers and we're writing stuff, Sean C Davis: Mm-hmm. Mm-hmm. Yep. Taylor MacDonald: The human brain language, the way we talk to each other, like this conversation here is very fluid and organic. It is not structured, right? And until the advent of natural language communication with computers, we have been forced to reconcile that reality. We'll parse content and we plug it into JSON blobs or we stick it into database tables, you know, and these things literally just serve to take our ideas and make them manageable. so that we can pull that out, the structured data, say, well, I only need the date value. And because it's strictly formatted as a date, I can turn it into whatever string format that I want that it comports with the design system or whatever I'm trying to render on the page. My argument now is that the CMS is effectively obsolete. as a function of this thing, this ability to talk, right? And a lot of people in this day and age, we're talking about AI. It's brilliant. There's so much knowledge attached to these systems. We can just say, hey, tell me how to repair the electrical component in my daughter's hoverboard or whatever. And it can give me really clear schematics. It can give me directions to A, step B, step C. And that's fascinating. That's amazing. You know, the knowledge ⁓ corpus that exists out there is really, think, what takes most people by surprise, right? And then when it is coupled to generative output, you're like, holy shit, man, this is the most incredible thing ever. That's not the thing. And my argument has been for some time now that like the biggest advent, the biggest benefit with this advent of this technology is being able to reduce that friction. If I can just talk at my computer, I can bring my typos, can bring my misstatements, my spellings, my humanism, right? The things that make me inherently fallible and human. are no longer blockers for me to interact with this technology, right? What do you think about that, Sean? And how has that influenced the way that your workflow goes today? Are you using voice automated tools or any sort of other integrations in your daily workflow today that have reduced that frictional barrier? Sean C Davis: Yes, yes, 100%. But I think you hit on something there that sparked a thought of, okay, you were talking about structure, right? And that one of the reasons CMSs existed was to put content in a structure. I think that enabled, ⁓ That was the mechanism that would enable stakeholders to build things that fit within the system. Because there was no other way to do it. The system was the code. This was our big thing at StackBit. You had to keep the code and the content separate if you wanted to maintain that system. Because as soon as you're like, Taylor MacDonald: Mm-hmm. Sean C Davis: I mean, you've seen it, right? Like the open WordPress fields and you're like, actually, I'll tell you what, like 15 years ago, I started blogging. And for the first decade of having this blog, my most popular post was how to add JavaScript to SharePoint pages because... People wanted to be like, I wanna add this little bit of functionality to a SharePoint page, and I guess I was the first person who wrote down, like yeah, you add this element, and then you can hack the hell out of this page. ⁓ I think to your point, it's removing the friction, it's fixing the errors, and something that goes hand in hand with that is that... Taylor MacDonald: Hilarious. Sean C Davis: If we build, it's kind of like back to my point of that middle layer and that set of rules that if you build that well enough, you also equip that agent with. way more freedom and creativity. And so we get to that point where we're like, actually, yeah, maybe it's okay if you build your own components as long as you're adhering to A, B, and C. So I mean, I don't know if I'm exactly answering your question, but is that, maybe I should turn it around on you. Do you feel like we're in a place where you... Taylor MacDonald: Mm-hmm. Sean C Davis: It would give you less anxiety to start to give more creative freedom to your less technical users if you can have the right rules in place. Taylor MacDonald: I mean, every, every client's different, right? Like I agree. Yes. Yes. Most of our clients want creative control until they get it. And then they either don't. Sean C Davis: You Taylor MacDonald: don't take advantage of that creative control. mean, countless clients we've implemented CMSs. And I will also argue that like for any website development project, mean, these days, I don't know, average website for a, you know, corporate entity, professional presence, you know, full end to end, you know, execution, considering things like accessibility performance, SEO, you know, obviously the CMS integration. would argue that like 50 % of the total cost of that. implementation goes to the CMS. I would also argue that probably 50 % of our clients don't even get in there after we launch it. Like, they're paying for this. But either knowledge leaves through natural attrition, right? Like an employee leaves and they're the ones that were trained in the CMS and nobody else knows how to do it and they just kind of don't know. And so either you have somebody who, you know, figures it out, right? I mean, at end of the day, it's not rocket science. Sometimes CMS is, you know, the implementation of CMS is very dependent upon, you Sean C Davis: Mm-hmm. Taylor MacDonald: your philosophy, what you're trying to accomplish, your prior experience, what necessary customizations the client requires. But ⁓ it's not uncommon for a client to call us and be like, hey, I don't even know how to update my CMS. I don't know how to update my website. Can you please just do this for me? And so they end up paying us per hour to go make content administration edits after they've made this huge investment on the CMS implementation. And so for years, I've had this argument that Sean C Davis: ⁓ man. Taylor MacDonald: Like hey, dude, you don't even need a CMS. Like if you want, let's just come up with a creative, financially viable approach that works for both us and you and we'll just handle it all for you end to end. Now, I've made that pitch to everybody for the last decade and I only have like one client that was willing to do it. I do think it served them well because you know, they need to run their business. Don't worry about getting in and finding the right field or finding the right form or you made this change and it rippled across. And over here on this other page, now that button is broken. because you didn't realize that you were using the same button that has been an instance on both of pages. And so I feel like I'm meandering on my response to what I will tell you is getting back to that prior statement of how do you reduce friction on behalf of the users, whether in a design sense or just a general structured data and data entry sense. Natural language has been the biggest change to my workflow in all of this stuff and and and by that I mean I'm using a tool called whisper flow which you introduced to me I started with super whisper because the syntax guys were talking about super whisper and I'm like, my god This is incredible. And then I found whisper flow and I found whisper flow to be significantly More sophisticated more mature better output better consistency and basically what it is Sean C Davis: Mm-hmm. Yeah, me too. Yep. Yep. Was it just that it ⁓ was like the structure of it or that, ⁓ was that it translated it better? Taylor MacDonald: Yeah, so I got better translation. I got more consistent like execution across. I have like a bunch of different machines that I use, you know? So like I would have to install it on all these different machines and it would work here, but it wouldn't work there. That's an Intel Mac. This is a Silicon Mac, you know, that kind of stuff. And so there was just like weird fragmentation nuances, I think that were tripping me up. But then beyond that, Whisper Flow, and let me actually clarify for the listeners what Whisper Flow is. Whisper Flow is a utility that you stall on your laptop where you hit a hot key and you can just Sean C Davis: Mm-hmm. Taylor MacDonald: talk. You basically put your cursor in any input field anywhere on any application hit the hotkey and you're like yo I want to send an email to so-and-so do the thing and yada yada. I don't I know you've been using it. What's your experience and did you have the exact same kind of revelatory epiphany when you first clicked that little hot, like it's awkward for the first week. You know, your wife's in the other room you're like, please send, you know, let's reply to Becky and do this, that and the other. But in the other room, your wife is like, what? I don't know if you have that experience or not, but. ⁓ Sean C Davis: Mm-hmm. Hehehehehe Yeah. My family, my neighbors, think that, you know, Cincinnati isn't the most technologically advanced city. I think I'm the crazy person in my immediate. vicinity because I'm always talking out loud. I got this co-working space and I'm glad I got the closed door office because I'm just like I'm just talking all day and people are like this guy's insane he's not even on a call right now. Yeah so I use that all the time. The interesting thing back to the friction point is so like that removes friction but it so so like I talk to I won't talk into the the computer like I do like I'm talking now is the ums and ahs and the wait I mean in trip over myself and whisper flow fixes that which is awesome and and that is particularly well suited for ⁓ emails and slack messages and output that some that another human is going to read ⁓ if I'm gonna go Taylor MacDonald: Mm-hmm. Sean C Davis: long because I think it has a max of five minutes or something like that and if I'm gonna go longer and care a little bit less about that say I'm just gonna take a walk and try to plan out the bigger and more involved feature that rather than going back and forth in shorter statements I'm just gonna just brain dump I use this tool called voice notes voice notes comm and that Trans, its transcripts are, they're very literal. It's like the Mac dictation. So the ums and the ahs and the repeated words are all in there. But you drop that transcription into a chat with an AI agent and it doesn't care because it does the work that Whisperflow would have done. So I think it just depends on like, what's your medium and who's receiving the output and how long is it? Taylor MacDonald: Yeah. Yeah. Well, I think, I mean, you just zeroed in on the one, it's resilient, you know? Like one thing, I can have a conversation, you and I have a conversation and I'll be like, ⁓ what was that one tool? And I get it wrong. ⁓ I've noticed that with these AI agents that like, if I make a mistake, often, so we've got an employee whose name is Ruto, R-U-D-O, you know Ruto. ⁓ Sean C Davis: Yes, resilient, yes. Taylor MacDonald: Often when I'm saying his name into some of these tools, will misspell, it will throw a D instead of a, or sorry, a T instead of a D, that kind of thing. ⁓ But it doesn't matter. Like the agent still is like, ⁓ yeah, like it's fine. It's totally self healing in a way that like, if I make a mistake, And I found this too, you I used to like wear out my backspace key, you know, because I'm a terrible typer. never learned. never like actually learned. It's sort of embarrassing. I've been in technology for like almost 30 years now and, and I'm, I'm still like, I mean, I can move fast, but I'm not, don't look at my hands when I'm typing in front of you, because it's embarrassing for both of us. That's, that's it. I don't worry about that anymore. You know, I used to have this like oversensitivity, like especially for client communications or Sean C Davis: Yeah. Hahaha! Taylor MacDonald: internal Slack communications. I'll send you a note, there'll be a typo, I'll command arrow up, fix it, hit enter, right? And I think that that, we talked over the years, Sean, you and I have talked about the cost of context switching, there's been research studies on this, that even those little types of things, those rote level tasks. These are things that we are cognitively doing that cost time. They cost money. They cost energy. They cost, you know, computational cycles in your brain. You you type the sentence, fuck, there's a typo. You arrow back. You know, I've gotten so good with the commands because I do so many typos that like I've learned if I hold option, I can arrow like by word boundaries and I've, know, that kind of shit. If I, if I hold command and option together and hit the arrow, goes to the front of the line. It's, you know, stuff like that. I've learned these little memory. muscle or muscle memory style tricks to help me navigate that faster because what I'm really doing is paving over my my own ineptitude, my own problems, the fallibility that I'm bringing to this as a consequence of being a human. Man, AI has totally freed me from that. Like I don't care about typos anymore. I mean, assuming I mean, they don't I don't have like raw stuff going directly to clients and I always review you gotta keep human in the loop. But like that is a game changer for me. you Sean C Davis: 100%. 100%. Yeah. Yeah, so we're gonna build something, yeah. Taylor MacDonald: Yeah, yeah, yeah. So, all right, to bring this back, right? So we've started this conversation about everybody hates their CMS. The biggest problem with the CMS is the friction. It's that you're putting a bunch of little boxes in front of my ability to just communicate. Human beings are organic. We just say things, right? So, did you ever watch Minority Report? Have you ever seen that movie? Yeah, yeah, it's old. I'm old. Like, don't judge me. I'm old. If you couldn't tell from all the gray. Anyway, Sean C Davis: ⁓ it's been a long time. Yeah. That's happening here too. Taylor MacDonald: It happens to all of us, unfortunately. Get ready. ⁓ So anyway, ⁓ in Minority Report, Tom Cruise plays this, ⁓ he's like a police officer, I think. He like runs this division in the LA Crime Department, maybe, I don't know what city it is. Anyway, it doesn't matter. He runs this police department around what they call pre-crime. And it's the idea that like they have these kind of clairvoyant, ⁓ you know, ⁓ these three women who are like psychics and they can see into the future and broadcast their images on a TV screen. All that's fantasy and kind of crazy. But one thing that he does in this movie that I thought was revolutionary, I think it was a Spielberg movie, so it's not that surprising, but like he stands in front of this wall of computers and he just talks. He's like, hey, pull this up, do this, do this other thing. And it's like this visual example of... of this exact same concept, like being able to speak your words and have the machine react accordingly with enough context to make it useful. I remember seeing that and just being like, ⁓ my God, that's the future. And then there was another movie called Her by Spike Jones where Joaquin Phoenix would talk to... ⁓ Sean C Davis: ⁓ yeah, yep, yep. Taylor MacDonald: Scarlett Johansson who was the computer and ⁓ and he is fascinating He ended up falling in love with her and it was like this book spoiler alert. It's great movie that go watch it Anyway, I That it but it but it was all centered around this voice communication It was it was the computer meeting us where we are in everyday life And that's here. It's here. Right? So imagine this imagine if you didn't have CMS Sean C Davis: That's right, yes. Mm-hmm. Taylor MacDonald: Imagine a world where the CMS was that thin layer between you and the architectural constraints that have been implemented on the front end. And you could just hit a button and say, hey, I want this over here. change that headline. The H1 isn't doing its thing. Like I need a different call to action. I want to build a new campaign page. It should look kind of like that, but I'm going to update it with this information. Maybe drag an image in. Here's my hero image. Here's you know, content. ⁓ I want it on this URL because that's where my campaign is driving traffic to. That seems compelling to me. What do you think? Sean C Davis: Yes, to me that brings it back to, you were talking about trying to sell your clients on, you don't need a CMS, we'll just do the content updates and you will spend less money over time. And you've proved that out with the one client. ⁓ Taylor MacDonald: Mm-hmm. Sean C Davis: That's how that client is, from their perspective, that's exactly how they're making changes, right? They call you up on the phone and they're like, move this, move that, do this, and you do it. what we wanna do, I think, is build the thing so that you don't have to take those phone calls. They can do exactly that. See it in preview, iterate on it, and then they're in full control, essentially. Taylor MacDonald: Yeah. Bingo. Completely agree and control is the operative order, right? If I ask myself why didn't the white glove content ⁓ management approach work for most of our clients or why do most of our clients seek a CMS even though only half of them actually take advantage of it to the full extent that it provides? It's all about control. We're control freaks. Humans want control. And rightly so, it's your business, it's your brand, it's your messaging, it's your initiatives that you're trying to bring to the web. So it's not lost on me that when I pitch a client, like, hey, dude, I got you, don't worry about it, just send me all your edits and I'll make them. And we'll own the performance, we'll own the SEO, we'll own the security, we'll own the stability, we'll just, it is a fully managed white glove solution. Sean C Davis: Mm-hmm. Mm-hmm. Taylor MacDonald: that the biggest barrier for entry or the biggest hurdle, know, cognitive hurdle for our clients is that thing, that lacking in control. Well, if I send it, you you might not get it up in 72 hours. And that could be a problem because I've got a deadline tomorrow, it needs to be done. Or I just don't like it. Or I've got 20 stakeholders that need to review it and they've got a thousand updates and I don't want to call you a thousand times because it's going to be expensive, right? I think... Sean C Davis: Mm-hmm. Mm-hmm. Taylor MacDonald: Yeah, think, think we're, how do we, know, using this technology, I think we're finally in a world where we can reduce the friction, give you the control, but then also build in those constraints early on so you don't shoot yourself in the foot. Sean C Davis: Yep, yep. I think that's spot on with what I've been thinking about and my brain starts going down the rabbit hole of like, ooh, all these bells and whistles we can add to make this super effective for folks. it feels like, yeah, what are you thinking? We should spin up a little MVP in a week or so, see if we can do it. Taylor MacDonald: Yeah. Yeah, I think that would be fun. in general, know kind of the general idea behind this podcast, behind these conversations that we'll be having over the course of the next few months is how can we take a concept like this, which is kind of interesting and lends itself well to the existing tool set. How do we build that? Like, what does that actually look like? And then everybody that's listening or following along come with a journey on us, right? Like if you're new to product development, you probably learn a couple of things. If you're interested in the concept of reducing friction, if you're interested in the concept of killing your CMS, I have long argued that, well, the arms race that is like headless CMSs in this day and age is obscene. Like in order to run an enterprise level decoupled CMS in 2020, Sean C Davis: Mm-hmm. Mm-hmm. Taylor MacDonald: it's almost an annual salary for some. It's absolutely out of control. And in so many ways, and I will tell you as an agency, we have this major problem where we'll pitch a client. Sean C Davis: Yeah. Taylor MacDonald: But given the volatility and the pricing and like continued inflation and I get it. I mean, the shareholders need their, their dividend checks or whatever, but we'll pitch a, we'll pitch a client on like, Hey, go use this not to be named CMS. It's going to cost this amount for the next year. And then they jack up the prices by the time, you know, it takes, we've got a six to eight month sales cycle. Right? So like at the end of the day, I don't have the guarantee that the pricing is stable enough because it's certainly not been, I've often argued that like, Content management as a service is about as I don't know problematic of product to be priced as is like healthcare. It's like the only industry that jacks up the price and reduces the entitlements beyond healthcare. It's bizarre. I don't know why we're doing it, because we all hate it, right? So anyway, ⁓ given that motivation and some of these passionate arguments here, I think there is a world where we take these concepts and we build a prototype and then we iterate on it. And some of the value statements here are how do we keep the cost down? How do we make it accessible? What's our posture on open source? What does open source even mean? I'm sure we have some listeners here that maybe are kind of trying to quantify what all this. Sean C Davis: Mm-hmm. Taylor MacDonald: talk looks like, how does it translate? know, if we're going to build a product and then turn around and follow the healthcare model of pricing, I don't think that's very helpful. ⁓ You know, so things like vendor lock-in are important, things like flexibility, maybe today you're using Claw, tomorrow you're using ChadGBT, interoperability of models to do the inference layer. I mean, these are all important considerations that I think we should, you know, use the next few weeks to kind of model those out and come up with a, with like a real roadmap and just run. Sean C Davis: I love that because we can kind of be building on the fly, know, share a little bit about what we learned as we built each week. And maybe there's a theme. I mean, I think it'll happen organically, but I think it's not just necessarily we're demoing what's been done, but there's a lot that goes into building a product even when you're not writing code. so that'll be, I'm really excited to explore that. Taylor MacDonald: Yes. Awesome. Okay, ⁓ well that's great. Sean, thanks so much. This has been really fun conversation. I've really enjoyed kind of unpacking this stuff or at least formalizing it. on a recording, because I know we've been talking about this crap for years. ⁓ But yeah, so the podcast is Blood, Sweat, and Tokens. We are building real products using modern AI tooling, new architectural principles, and really trying to revolutionize the way that we think about how humans interact with software and what the potential benefit could be for the community as well as the industry. And if we can kill off a couple of, you know, Sean C Davis: Mm-hmm. Taylor MacDonald: exorbitantly expensive CMSs along the way. All the better in my opinion. All right. Sean C Davis: Love it. Love it. All right. Yeah, let's ⁓ let's get into it. But yeah, great chatting with you today, Taylor MacDonald: Okay, sounds good. To be continued, we'll be back next week and we'll, you know, kind of start that process and start chipping away at the problem statement. Alright, thanks, Sean. Bye. Sean C Davis: Fantastic. Thank you, see y'all.