Why Design Still Needs Friction When AI Moves Fast

What We Need to Grow, Episode 078: You Still Need the Friction. A Founder Talk conversation with Matt D. Smith. Cleaned for readability; the words are the speakers' own.

A Conversation with Matt D. Smith, Founder of Shift Nudge

Cleaned and publish-ready. The full Riverside take, 00:00 to 45:07.

Caleb Pedosiuk: All right, Matt, let's do it. So we met a couple of years back. We were in Kelowna, a beautiful part of the world, and man, I so appreciated getting to chat with you. Some people, you have some idea who they are before you meet. You're one of the instances where I got to meet you, and then I found out after how legendary you are in the design world. I think even Kanye at one point reached out trying to get you to help work on some stuff. I think you may have done some stuff with him, or not. I know he's a very interesting, dynamic figure, and there are different eras of his work. But from what I've come to understand, people have looked to you as a leading perspective in design. So by all means, please share a bit about Shift Nudge, what's going on, and how it's going.

Matt D. Smith: Cool. Well, I appreciate that. It was a pleasure to go to Kelowna, that whole experience, because I did not know a single person. It's very rare that I go to a conference or a big event like that where I just don't know anyone at all. So it was nice to get a chat with people like you. You bond quickly over similar ideologies. And I think any time you're going to events that are very focused and specific, you get a little built-in values from everyone you're talking to, so that's always nice.

It's funny, because I was just meeting a new designer who's trying to get a foothold and looking for advice. He's about to move to San Francisco and lives in my town. He was just over here yesterday, and he was like, "What does it take? How did you become this legendary designer?" He kept using all these big words, and I was just sitting there like, man, I'm just some random dude who grew up in rural South Georgia. I just started designing and I never stopped. It wasn't some big master plan to become this legendary designer, and I don't necessarily think that way either. I'm happy to hear someone say that, but at the same time, I feel like I've always been curious. I've always been passionate about the things I was creating or building, whether that was a kid drawing Mortal Kombat characters or Bart Simpson, or getting older and figuring out creative ways to design interfaces, or the agentic experiences we're all working on now.

So my entire career has been designing. I went to school for design. Before that, I was very creative and built things with my dad in the workshop, working with wood, and did a lot of home renovations. I just designed, designed, designed. I was very fortunate, because a lot of people go to school or university to study something and then their career takes them in another direction. I stumbled into graphic design and continued doing design after school, and I've done design for my entire career: smaller jobs, then agencies, then startups, then building some of my own products, and continuing to work on a lot of different products from a user experience and user interface design perspective.

Now most of what I'm doing is teaching other people to do that same type of design, and that's what I've really been focused on since 2020. You alluded to Shift Nudge, which is my design training brand. Through that, I've been teaching designers interface design and introducing some new AI methodologies. That's the quickest way I know how to tell my backstory as it relates to design and what I'm doing now.

Caleb: And if I'm not mistaken, you've had something like 5,000 designers come through your program?

Matt: Yeah, probably five or six thousand or so. Some come in as individuals. Either someone wants to get into design, or maybe they're four or five years into design but just aren't as good as they want to be, and they want to figure out how to take their work to the next level. They look at all these other designs and they can notice good design, but it's hard to know what's wrong in their own designs, why it's off, and how to fix it. So that's the entire goal for me: helping people see what's off, understand why, and know how to fix it.

I've also been fortunate to have teams from larger companies. Someone who is the design ops director will bring in 10 or 20 people off their team. A lot of big brands have put in designers, sometimes project managers, sometimes developers. It's been a big gamut of different types of people who just want to learn design and get really good at it. And now a lot of people are building with AI, and they're like, "I can build things and it's cool now, but I just want to learn more about design." So I'm starting to navigate that as well, since it's a little more challenging to get someone to slow down and learn some of the fundamentals that will make you much better when you're going a thousand miles per hour.

Caleb: It's funny, the paradox right now is that yes, at the click of a button you can do 10x, 100x, 1,000x the work you did before. But the flip side of that coin is that now there's 10x, 100x, 1,000 times the amount. It's like deflating a currency with a money printer. What once took one unit of energy to produce one output, one unit of human energy will now be producing a hundred outputs, and that's across the board. The water has lifted the tide across the board.

So now more and more people have a heavier callus in their resistance to responding to things, in the way they scrutinize content. The BS filter is there. The amount of noise out there has been cranked up. So the question is how to find the signal of what someone is actually looking for, whether it's a real problem they're trying to solve in their life, or something that resonates with them. Certainly from a brand standpoint: if a brand is more than the actual thing, more than the brown sweet cola or the sparkling water, then what is that thing Coca-Cola or Pepsi has tried to evoke in someone? If it's more than the caffeine and sugar people reach for in a Red Bull, then what is it? And how does a brand take that and craft experiences that not only interrupt someone in a social media feed to get their attention, but then, when they move over to some ecosystem, their website or whatever, draw them in? How are you distinguishing yourself, evoking that feeling and connecting it to whatever you're doing, product, service and so on? The importance of design, in an era where design seems so easy, has actually become the need for good taste, for things to be crafted well, all the more.

Matt: Yeah, and I think it's interesting too, because there are a lot of times when really good design isn't necessary. You've probably experienced this yourself, where someone finds out you're a designer and they're like, "Yeah, I've been needing to get a website. I've got a pool cleaning business and we need a website." Most of the time I'm just like, bro, just go to Squarespace and set up a template. In 30 minutes you'll have a website. You'll spend a hundred dollars. You don't need to spend thousands of dollars on me to get a website when there are a lot of templates that are good enough already.

And now you might not even need to do that. If you're willing to get your hands dirty a little bit, you can just say, "Hey Claude, I need a website for my pool cleaning business. Connect to GitHub and deploy it." Squarespace or any of those others are still good. But when I was early in my career, I would jump at opportunities like that and meticulously design it, and that might not have been the best first step.

So in this whole process where we can now go a thousand miles per hour, I'm constantly pondering it. I believe in good design. I want to do good design, and I want to teach people to do good design. But I'm also trying to map out the best places to insert the good design skill set. Sometimes it is good to have that mindset from the very beginning. But depending on what you're building, it could just be, "We need something that works, that solves this one problem with somebody calling on the phone, or pulls the transcript." It's kind of ugly. I've got ugly systems that I use that I just haven't designed well, because I don't need them to look beautiful. I would love to spend the time, but you don't always have the time.

So I'm trying to find that good balance of when you spend the time on this perfectly crafted form field, or interface, or brand style guide. Because AI can do so much now, we can't necessarily spend so much time on every tiny detail. But if we can get it really great from the very beginning with brand or design direction or style, and someone is in charge of making those decisions, then I think that can also serve as a new filter for AI to keep building in the hands of everyone around it, with style guides. But design systems, style guides and brand guidelines only work well when someone knows how to use them. You can still mess things up with those things in place. It's very contextual, so it's difficult to make blanket statements about design sometimes.

Caleb: We had one client for about seven years, and we worked on their integration strategy as they were acquiring companies and merging them. They would acquire half a dozen companies a year, and we would take everything from white papers to solution briefs to webinars, you name it, and integrate these companies, primarily product-driven companies, into the new parent brand, and make sure every i was dotted and every t was crossed. During a merger there are a lot of redundancies, and the precision and effectiveness of how well a merger was going was very visible. It was like a canary. If all of a sudden assets are going out that are these weird Frankenstein things, that's a sign. It has to go smoothly. The benefit is that when the teams see it running smoothly, when directors and stakeholders see it, they think, wow, this seems like it's going really well. That's an instance of design.

But to your point about wondering where to use it, you have to have seen the SpaceX engine. Was it the Falcon engine? I can't remember what engine it was, but it shows the first version, version two and version three. The first version has wires all over the place and looks very much like it's been put together in some garage. By the third one, the engine is just beautiful and minimalist. I think that's the idea: make it work, and then you can make it look better later.

I've wondered about that specifically as it relates to dashboards right now, with people spinning up dashboards in Claude. You have all these assets that are basically the next generation of Google Docs, a lot of artifacts being put together that reflect frameworks and so on. What do you do with them all? Now there are dashboards being created so that you can index them. I've been wondering if a new era, a new wineskin, if you will, for this era is the idea of more compostable dashboards. We build them with the intention that this is going to live for six months, get used really well, and then probably return to the pixels it came from, as opposed to trying to perfect it.

Matt: Yeah. Sometimes I think of it as scaffolding. We're building the thing that's helping us build the thing. When you're building a really nice building, you've got this junky scaffolding on the outside, but it's very critical. We can't reach where we need to reach unless we meticulously build and secure the scaffolding, but we don't need it to be fully polished. It can be kind of beat up. Things are going to hit it.

And some of the things we're building with AI, some of it is just because we can. The same way we can option-drag things across a canvas and play around with design, we can also prompt 20 versions: show me 20 versions of the same component, only adjusting the negative space, or show me 20 versions only adjusting the colours. It's just so easy to produce that. So it's about knowing where you are on that spectrum, whether this is scaffolding or the real thing that needs to be finely tuned because hundreds of thousands or millions of people are going to use it. I don't know that anything has rapidly changed from a crappy-prototype-to-finished-product point of view, other than that maybe you can get closer to a finished product a little faster. It depends on the product, and it depends on the person wielding the power creating the product.

Caleb: So what has it been like for you? If I get this timeline right, you're in Adobe, that whole suite, learning those toolkits. You see Figma come up, with some real strengths, pioneering. Then you see the entry of ChatGPT and Claude and all that jazz. From what I understand, you even educate when it comes to understanding Claude Code: lessons, prompts, helping shape that. What has that been like for you? Is it one big wave, constantly learning, or is it multiple swells?

Matt: I definitely started in Adobe Photoshop, and Macromedia Flash and Fireworks and all those things, before Adobe bought Macromedia, and then Illustrator and some vector-based tools. Then I got really into Sketch. Sketch had a run of two or three years where it was kind of dominating the interface world. Then Figma came along, and I was reluctant, even to use Figma, because it was web-based design. I was used to having my precious files, my PSD or my .sketch file, on my computer in my safekeeping. So it was a challenge to adopt a web-based design tool. Looking back, I think that was maybe 2017, nearly ten years ago. We were clinging to our local files and going web-based. And now we're going back to local files in a lot of ways, because that's what code bases are: living on your computer and pushing to the cloud.

So I feel like the only constant is change. If you focus on the fundamentals of problem solving and any other principles that don't change that much, think about them often, and hold your tools loosely, I think that helps. There is still some reluctance to let go of my manual movement on the canvas, because I still think there's value there. So now I'm asking myself how much of it is just my opinion that this is how I learned, so this is how it should be learned, which is not necessarily the correct viewpoint.

Why did this help me learn negative space or typography? Adjusting a type box, adjusting the line height, adjusting colours slowly but surely has a sort of friction to that learning process. I think the friction of the learning process is critical in a lot of ways. You see it looks pretty good, a little bit better, a little change here, and then, there it is. You went on this journey of maybe, sort of, kind of, and then yes. It's hard to reproduce that exact process when you're building with AI, because it goes so much faster. You can jump from point A to point B and generate a bunch of options.

I think you can get close, and I think there is a way to learn, as long as you're introducing the friction of needing to slow down and say, "Okay, here are 20 versions of something. Now let me take the time to go through them. Not three. Yes, number five is pretty good, but this might not work." You still need the friction. So I'm always trying to figure out which parts of learning design or doing good design are necessary, whether I'm building it this way, the old way, or both. I've determined for myself, and in some of what I try to teach, that a certain amount of friction is necessary. Even if you can build a million miles an hour, you still need the friction of slowing down and making that informed decision based on any number of inputs.

Caleb: What are your thoughts on the rise of MCP? Any platform that isn't allowing for compatibility through MCP runs the risk of becoming disconnected from the economies of exchange online right now. It almost invites a gap for innovation to spin up a version of what a company or product has. Higgsfield, for example, is shooting up like a rocket. Were they the fastest company in human history to reach... I don't even know what their ARR was. Something nuts. I'm not exactly sure what the amount was. I know Slack held that crown at one point. Look at how they're doing the Vox-style animations and text editing, taking things like your brand guidelines and your fonts, and letting you edit video from text prompts just by accessing folders on your computer. What's your take on some of that?

Matt: I think everyone will always want speed. If it's Amazon Prime, I'm going to choose that over Amazon regular, and if they could ship it to me in the next five minutes, even better. I think Jeff Bezos has said it before: he liked to focus on the things people are never going to not want, and speed is definitely one of them. Jason Fried from 37signals talks a lot about this too. I don't know if a lot of people know this, but I think Jeff Bezos has a tiny stake in 37signals as an investor, and because Jason is planted in the tech and design world, I've heard him paraphrase that as well.

Because of that, I find myself writing my own skill files. I've got an agent Python-writing skill. So it's like, "Hey, I want you to write some Python to access this API, because they don't have an MCP, and I want you to take all of this information, grab these photos and throw them in the newsletter, and just do that for me." I like it because some of the interfaces are just slow to use, and if you're very connected with your own agent, you can do it so much faster. It's hard to go back to doing something the very slow way once you've experienced the very fast way. If you've ever started using MCPs and tapping into things through an API from your Claude Code terminal, you're like, my gosh, I didn't have to do that. I was on my phone, drinking coffee outside, and I spun up this entire whatever it is, because I connected to everything.

I just think there's no way we'll revert to saying, "You know what? We need to do this by hand. I need to go in and physically click that drop-down button, then go grab that photo, then do this and do that." And video editing is one of the most time-consuming things you could possibly do in a creative field. Motion graphics and video editing are very, very time-consuming. If you have the ability to speak to a tool that can go do it for you in a quality way, and you can direct it, "Actually, no, don't do that. Cut it back here 20 seconds," that's a lot faster. Who wouldn't want to go faster, and who doesn't want it to be easier?

There are a lot of people with very political opinions about AI and very strong stances against it, but that's just where we're moving. There's no world in which this does not become a dominant force. I can either decide, absolutely not, I'm going to be a farmer from now on, although I'm sure a lot of farm equipment will probably introduce AI, crop monitoring and things like that. I just don't know if there's any escaping it, and I don't have a strong enough stance against it to not fully embrace it. Technology usually wins. People made fun of the internet. People made fun of the television. And people made fun of AI when it was first here: look at the Will Smith video with the spaghetti fingers, AI sucks at generating video. Now I can't even tell you how many Instagram Reels my wife has sent my kids, and they're like, "Mom, that's AI." And she's like, "Is it?" So I think it's inevitable. If you are not trying to figure out how to incorporate that, and how to make it easy for your agent to work with my agent, to work with my MCP or my server, I think it will not go well.

Caleb: So let's do a thought exercise. Let's pretend that anyone running point on a brand, anyone creating content, can now snap their fingers. They've got the magic prompt that populates a number of compelling motion graphic pieces: shorts optimized per platform, long form, short form, mid-form, the right dimensions. Everybody who needs to be tagged is tagged. The captions are right. It's not AI nonsense, it's just simple enough, it's signal, and it's visually interesting, across the board.

Pause on that thought experiment for a second. If part of your Claude Code course, your education in Shift Nudge, helps people understand how to talk to these models to optimize these things, I think that's so helpful right now. But that's just on the margin. In this thought exercise, if I speed up the tape and look at where this board game goes, if everyone is able to optimize for it all the time, the tide has risen again. Let's say back in the day, having a top hat made you stand out, and people said, "The person with the top hat, let's buy our newspapers from that stand, because that's the smart one," or "that's the politician to vote for." Then everyone's got top hats. Now top hats are no longer top hats. They're just hats. In that sense, I could see someone saying, "I don't have a top hat. I do it all by hand. Fully made by hand, certified." Okay, I can see that, the way someone would handcraft whatever they choose to. There's some neat traditional stuff. I think China does a bunch of that with the content they produce, holding to the traditional ways and capturing them in a technologically advanced way.

But to me it comes back to this, and you mentioned farming, which I think is interesting. I regularly think: what is the point of all of what everyone's doing? That's actually part of the reason for this conversation series. We'll hit 80 episodes this coming week, in about three and a half months, and it's talking to founders and asking: what are you working on, who does it help, why does it matter? To me, the centre of it is what meaningful problems people are solving. A lot of design, a lot of these things, are a means towards that. When we come alongside a company to help them, or when you come alongside a company, it's made me ask: why should someone care about the motion graphics? What are we doing? I don't mean that in a critical way.

In a way, I hope this is a signal I'm sensing that others are sensing too: where can we shine? How can we take this increased momentum we have across the board and find ways to help? Maybe, to use farming as an example, we can find some of these regenerative agriculture pocket businesses that couldn't get their signal out before and help them amplify that signal relatively quickly. People say, wow, nutrient-rich food. Demand increases. People find them for what they need, and maybe it helps with deficiencies they have. It helps those businesses grow, creates more demand, and helps more businesses like that do the same. And not just in farming, but in all kinds of cottage industries and bigger industries, where this is applied to something more than Gen AI clipping from popular shows, attention hijacking, a business model where the more it rises, the less it helps the people using it.

Matt: I think a lot of people just accept that optimizing is valuable. We're optimizing our workflow, streamlining our processes, everything is optimal, AI agents, optimum metrics and rates. But the ultimate goal isn't necessarily to optimize, to increase our efficiencies. There's the book, The Goal, about the theory of constraints. The guy in the factory is struggling to understand these really expensive new machines that are supposed to optimize their efficiency. Then he meets a guy in a bar who's supposedly some consultant, and the consultant asks, "Are you making more money?" And he says, "No, we're increasing our efficiency." "Yes, but surely your profits have increased as well." "Not necessarily." The whole book talks about how increasing efficiency is not necessarily helping you make more money.

Back to what you're saying: if you run with that thought experiment, where everything's clipped, everything's cool, everything's motion graphics, everything's optimized, short form, long form, that's not necessarily the end goal. Ideally you have a goal where you are affecting someone in a positive way and providing value, in exchange for their attention, or in exchange for a good or a service, or vice versa. If you squint your eyes, step back and look at the future timeline, it really makes you ponder the deeper questions: okay, what are we doing here, if we're just moving a million miles an hour and everything is optimized?

Maybe that's where a lot of the ideas Elon puts out come from, that we're not going to work in the future, everything's going to be taken care of by machines and we're just going to hang out. That doesn't quite sit right with me either, because I think we were created to work. I think that's where a lot of the tension is. If I was created to work, and I believe it's my duty to sweat and to produce with my hands or my mind in some way, and that's part of being human, but it feels like AI is sort of undermining that, of course it's going to create tension. So it gets to a more important question: what am I doing, and why is it important? I think those are great questions, the ones you proposed in the mantra of the podcast. That's a good question to ask.

Caleb: As you're saying that, I'm picturing it. Recently I've been chopping wood with just an axe. We've done it with a splitter before, and it's fine. It's a very different experience with the sound of the splitter going, electric or diesel. It's very efficient. But swinging an axe and chopping wood does something to the body. The testosterone goes off the charts, apparently. But there's something in that process, going from a signal, "here, I'm going to do this," keeping some form, exerting it, the physics meeting it, the sound, everything. There is something happening. After a few minutes of it, I'm like, this is something.

I've also noticed it with the guitar on the wall. I'm by no means some savant when it comes to guitar, but I know enough to tune it. I'll sit by a fire, and you hit a string, and the sound, the feel of that string, or a couple of simple notes... I'm telling you, there's something in here that, when your fingers move it and that sound moves out, is in sync, is amplifying, is tuning something up. So what if Optimus is just splitting the wood? What if a bot is just playing the music, a symphony beyond, whatever kind of symphony you want, here it is? But it wasn't about the complexity.

Recently I've been sitting down at a little electric piano we have, and I'll put on some instrumental, classical sort of music. If you put the JBL under the piano and turn up the volume, you find a note you're feeling, and you bring it up and down with the volume. Sometimes it's just two or three notes. Somehow an inside part of me is joining with what is happening. Whether prayer or worship, as a Christian, you're often in settings where you're expressing gratitude, being so thankful to be alive, thankful for existence. And when you come back to it, part of what we're created to do as created beings is to work, to express, to do things. I a hundred percent agree. Even if all the food were provided for, I told my wife this year, if we didn't get a single unit of food out of gardening, I think all of the work is worth it, just from all of the other benefits that come from the soil, the rain, the weeds. There is something really beautiful about being in nature in that way.

Matt: Yeah, I a hundred percent agree. You mentioned, "Here's a symphony. You want it like this? Here it is, instantly." And you're like, well, yeah, but I want to hear your crappy version of the song on the guitar. I want to hear you sing it. There's something different about someone sitting down with a guitar and playing with emotion, with skill, and even with some mistakes. That's just human and real, and it can't be replicated, even if the exact same song is played extremely well by a generated video of you. That's cool, I'll watch that too, but I'd rather make sure I saw the real event happen.

I think a good metaphor for this, for me, is photography. Photos at this point are very easy to generate with a couple of models. There's Donald Trump riding an alligator on a 747 in the clouds, and it looks very realistic. You could probably generate a photograph of me standing over there by my guitars, but it wouldn't be the same as taking a photograph. For example, we went on a big Yellowstone National Park trip a couple of months back. My oldest son was finishing some classes over the summer to get his certificates and finish the program he was in, so he wasn't able to come with us. We have all these photographs of our family in front of these big monuments. I could generate a version where he's standing there, but he wasn't there. That's just real life. You didn't see the grizzly bear. You didn't see the Grand Tetons. We felt the pain of him not being able to see it, and we talked about it. That's real, and those photographs are real, and you wouldn't want a generated version of them, because they're documenting reality.

If you use that metaphor of photography and what it represents versus generated things, there's a connection to reality that is still valuable. But that might not be the same for a bottle of lotion on an e-commerce website. Generate those. I don't need to know the model's hand was there; if the photo looks cool with the model's hand holding the lotion, that's fine. But you could also make a whole thing out of it: here's our documentary of us filming these lotion bottles, it's actually me here, no AI. That's also an interesting angle. So I definitely think it's forcing everyone to make sense of the world we live in a little bit more, because things are so prolific and profound and easy to do and easy to generate, but at the same time, they're not, in a way.

Caleb: It's funny, I was chatting with Graeme Barlow. I don't know if you've ever connected with him. A really awesome dude, one of my favourite people. One of the things he was talking about is how much of a bubble a lot of AI is. Depending on your circle and the echo chamber of what you see online, you might feel like everyone is so advanced and using it. Then you go outside that bubble and realize most people are drafting the odd email. They're learning to interact with it. A lot of people are so focused in their own ecosystem, a lot of it a legacy system of the internet and technology, and a lot of real businesses that solve meaningful problems and produce good quality services and products are completely removed from it, which is helpful. But I do like to race ahead in my mind, picture where things are going, try to get ahead, and see how we can benefit from it. Maybe creatives have some of that bent in them. I think part of the responsibility is to think through some of that scenario play and then try to line up our choices with it.

Matt: Yeah. On the flip side, we've been covering some of the doom and gloom part of it, but I've also been building a ton of things I would never have been able to build, exploring ideas I would never have been able to explore, and setting up workflows and automations I would never have been able to set up, using these tools. It's really exciting. It's really fun. If you're choosing to have fun, it makes things more enjoyable, versus doing things from a place of fear and anxiety, which I've also experienced. It's about choosing to see the good, the silver lining in any of this, every single day, every single month.

It's also easy to get online and read that a lot of the AI marketing playbook is "X job is cooked." Fire your designer, fire your developer, video editors are cooked. There's a lot of doom and gloom in the marketing cycle from these companies. But they're all just tools, and someone still has to use the tool. So it's been a lot of fun to avoid the hype of the fear cycle and just use the tool to build things I think are valuable, and to use it to help teach other people and show what's possible to someone who hasn't been using these tools: look, you can do this in 30 minutes when it might have taken you a week before. Let me show you how.

Opening up some of those possibilities, starting to use it creatively instead of resisting it out of fear it's going to take your job, opens up a lot more opportunities. We're in a very big change. The industries are all changing, with MCPs and agents and all these different things. Really, the only thing I think you can do, if you want to stay in the industry, is find a place to grab on, hang on, ride it out, and start playing with all the things you can play with. It is interesting, and it does make you question the work you're doing and why. But as long as you have a little clarity, it helps you move forward confidently.

Caleb: Totally with you. The Goal: A Process of Ongoing Improvement, the business novel that introduces the theory of constraints, by Dr. Eliyahu Goldratt and Jeff Cox. That was the book you mentioned.

Matt: Yes.

Caleb: Last question I love asking guests. If you had to put your finger on one thing, and I know it can be tricky, what would you say you need most to grow right now? Some people are raising money for something, some people are launching something. I'm not sure what's going on with Shift Nudge, but what would you say you need most to grow right now?

Matt: I'm very much a builder. I love to get in my workshop and just build and play and experiment, build, build, build. I've gotten into producing content and sending out my newsletters, but I will naturally do a fraction of the marketing compared to the amount of building. If I get into marketing mode and I'm trying to do all these random things on the marketing side, I start to feel odd. I want to get back to building.

Because of that dynamic, I'm still handling most of the marketing. I have a video editor who helps me produce content, and someone who helps on the operations side. But I find myself needing to focus on marketing, or rather, I need to reach more people and bring them into my world so that, if they choose, I can help them get better at design and building with AI. So the thing I need is to reach more people, simple as that. That's not always something I want to focus on, because I like to do these other things. I like to figure out new ways to help the people who are already in my circle, when I should probably spend more time reaching more people. So that's my long explanation of the one thing that could help me grow.

Caleb: Cool. Matt, thank you very much.

Matt: Thank you. Fun, fun chat.