Why AI Is Replacing the Agency Pitch Deck: Travis Weerts

It's so possible now to build all these ideas pre-AI. There was so much work that you did not communicate all of the depth of it to a client. To sit in front of a computer screen day in and day out building these digital tools, that's not a valuable use of human life. Really. And to have AI kind of step in and then take some of that stuff out, I think it's great for people. And to be honest, it just pissed me off.

Welcome to Beyond the Brief. On today's episode, I'm talking to Travis Weerts, a man of many titles, be it designer, developer, artist or even maker. Travis has built many things over his career. From AI radio stations to room visualizers, voice agents for airports, Australian open activations and even a running app that started as fitness motivation for Travis, but ended as a protest against Spotify's investment choices. Travis has made some incredible things, which is why I wanted him on the show. Because for years, a lot of creative work has died in pitch decks. He had the idea, but had to sell the vision to make it real. Travis argues that those days are over, and that he's seen a build first mentality, where it's much better as a sales tool than just the idea itself. In this episode, we talk about the future of agency work, how his role is changing due to AI, and what happens when the people with taste can finally build the thing before a committee has time to kill it. So let's get into it. Travis, how are you?

Good, man. How are you going?

Good. Welcome to the show. Episode three. It's good to have you on.

Good to be here, man.

I said this before to you, I feel like we should have crossed paths earlier. You seem like someone who's doing very cool things creatively with a lot of different technology. I guess the title that I'd probably give you from what I'm looking at from afar is creative technologist. I don't know if that's true or not, but it seems the kind of thing that you do. Yeah,

I think so. I hope so. Sounds about right.

That's good. So I guess for viewers, why don't we start with a little bit of a backstory, where you've come from, what kind of work you've done over your career, and we can kind of hone in on some of the brilliant things that you've created, of which there are many, after doing some research. And yeah, we'll just see where this conversation goes. Yeah, cool.

I guess, first of all, yeah, likewise, I don't want to rehash it, but I feel like, yeah, we should have crossed paths sooner. It feels like, yeah, you're a brother from another mother or something, working in the same space, same kind of, definitely thinking the same, similar ways. So it's good to finally, yeah, meet and have a conversation and catch up. Well, hello. Yeah, hi. I won't belabor it too much, going too much in my background, but creative technologist was my latest sort of job title. I never really thought of myself in those terms or with that title necessarily, but once it was sort of bestowed upon me by my elders, I was like, you know what? I think that's it. I've just always been into the tech space, design space, creative space, and I sort of had a little bit of both, I don't know, left brain and kind of both and trying to find, I guess, the juxtaposition of those things. You know, I sound like I'm going way back now, but even if like my earliest memory is like in this space is like spending my nights in Barnes and Noble in the programming section, like reading the like O'Reilly books and stuff, like trying to figure out how to build apps that I was deploying in like AOL. That's how much I'm going to date myself here.

Pretty Stack Overflow, you know?

Barnes and Noble was Stack Overflow, man. But yeah, I've just always had like a fascination for it and then decided to study computer science, even though I absolutely wanted to go to art school. But my parents, you know, looking back rightfully so, like said, maybe let's not do the starving artist thing, even though I'm still somehow feel like I'm doing that. Yeah, and then I freelanced for almost my entire life and got some amazing opportunities to do stuff. Just super lucky. Just, I don't know, being passionate about what you're doing and just saying yes to projects just leads to more work and more work. And there's nothing, yeah, complicated about it. And yeah, I did that for a really long time. That's a short story. Went from small mum and pop projects or indie bands all the way to some incredible opportunities never even dream that me as a little freelancer or whatever, get to work on.

So when did you arrive in Western Australia? How long have you been here for?

About 10 years, I think.

Okay. Yeah. Nice, nice. A little aside, but how do you find it to the rest of the world in terms of creative opportunities and creative projects and creative thinking?

I feel like I've gotten to do even more incredible projects. And since I've been in Australia, the opportunities that I've been given is just amazing. Really, I find it great, really. Yeah, great. I feel like WA kind of always gets this rap compared to like Sydney and the Eastern States and stuff. But I don't know, that's not been my experience at all. It's rather than creative.

Yeah, I think there's a lot. And you've had quite a career, I think. So as you're saying, you kind of started out freelancing and you've worked at different agencies around town. Yeah. I think you were at Dilate for a bit and then VML, if I'm right. And you worked at the start for a bit. I don't know, you might still be there. I don't know.

No, I'm now freelance. I've decided I've been working towards getting back, I guess, to my roots.

And is that just because it's about sort of creative freedom, you know, finding the right project and something that sort of feels right for you and where you can just dive in 100% or a bit more free? What is the desire there?

Yeah, absolutely. You just nailed it. Yeah. At VML, I was there for maybe five, six years and I changed names a couple of times, but it's not important. It's a little bit more, I'm not sure what you know about VML or not, but it's a global agency kind of thing. In that world or knew much about it. And coming into it, I definitely felt like a fish out of water a little bit. I was, you know, I've got my digital background and then going into sort of more classic, not classic, I hate to use the word classic because it makes it sound old, but traditional advertising style. Yeah. Where everything is campaign based and you're constantly pitching ideas. It was fascinating. It felt like you were living in an episode of Mad Men, even still today. Rapid creativity, right?

Go. Yeah. It's not that easy. It's really hard.

Except that last bit, like the go bit never really happened. It was more like, here's an idea. Okay, now let's go on to the next idea and the next idea and the next idea. Right, right, right, right. I, like as a person that likes to build and design and see ideas through, like that was just really frustrating, I think. Yeah.

Like. You're just kind of throwing these things out and not being able to see the tail end of it.

Exactly. And then it goes to some other team that, yeah.

Which is fine. It's how those sorts of businesses are set up to serve those types of engagements. But I guess as a freelancer now, you can, you can have the idea, you can craft the idea, you can deliver it. Totally. See the performance. Yeah,

And there's a joy that comes from making a thing and putting that thing out into the world. And I just miss that. And so, especially in this new world that's been emerging, I was like, it's so possible now to build all these ideas with, you know, little effort or less effort, I should say. It's not that like it's little effort.

I'd classify it as a different kind of effort.

Yeah.

It's still a lot of work, but it's just, it's deployed differently and it's amplified. You can get more done. Totally. That's really cool. So, you know, you're a, you're a builder, it seems. And, and I think product is something that probably comes pretty naturally to you then. So the idea of seeing things through and, you know, starting with an idea, shipping something and seeing it grow and improve is probably pretty alluring. You seem to have done a lot, you seem to have done a lot of that. So, I want to touch on that soon and start talking about some of your building products. But what I really want to touch on was something you just raised around the changing of the creative process where, you know, you're playing with these new tools. You've, you've gone freelance because you're sensing a sense of possibility. How did you, when did you first start playing this, this technology? So LLMs to boost your own sort of workflow. And what was that journey like to where you are now?

I can't remember the, like the very first time I can kind of remember maybe around the Midjourney. One point, the first version of Midjourney, seeing it. Yep. And I think right at that same time, I was also kind of seeing the, I guess, just immersions of more public facing AI products. I think there was, there was one from IBM that could read the sentiment analysis in a body of text. And then I was like, oh, we can, we can do really cool stuff with that. And then seeing the image generation, I just started playing with it at that point, like little weird ideas that went nowhere and just experimenting mostly just in the spirit of play. Really. Like, like for example, the IBM one, I started wondering if you could like, I don't know, build a body of text from a bunch of people and display the emotional side of that text in, in form of a, an interactive, like painting. Cool.

Where'd that get to? That sounds cool.

It, it, we, we, there was a project that came in at VML for Murdoch and they, wanted to display some of it, some of that thinking. It didn't end up looking like the visualization that I played with and built, but that idea went on to, they've got a new technology. I can't remember the name of it, a technology centre that they've opened, new building and they're putting in some displays. And that idea ended up being on one of those displays around a crowdsourcing, emotional analysis art. I can't remember what the name of it was, that they put on it, but yeah. Yeah. But it started with.

It's all the light of day.

Yeah. Hey, cool. Let's try this thing. And then, well, there's this idea that that might work for let's pitch that. Yep. Yeah.

Yep. Kind of seeing, seeing a twinkle of capability and then say, okay, well how can I use this and how can I be creative with it? Yeah. I love that. So what followed up? I remember the early days of Midjourney and it, it kind of blew me away. Like I saw the really early research papers from Google. I think it was the Google DeepMind where they had like, here's a cat. And it was just like a pixelated graphic of this is what a machine thinks a cat looks like. And I remember that absolutely blowing my mind. It actually looks like a cat. Yeah. It's kind of this fuzzy pixelated cat, right? And I'm like, this is incredible technology. Imagine what we could do with this. You know, and then Midjourney comes out and everything, you know, even that, that was really exciting. You can do a lot with this, but the quality, you know, it's really creepy looking. Totally.

Where did they get that? It was from that they trained it on.

Yeah. Yeah. I mean, that's a whole separate issue we can talk about, but yeah, just seeing the progression of Midjourney was amazing. I, just as an aside, you know, I was playing with it. I'm sort of a shareholder in which is menu Zen. And we were using it to generate food photography for our menu designs. Yeah. Because it was, it was much obviously, you know, startup, you bootstrapping. It's much more cost effective to generate photos. And the quality just blew my mind. I thought this is, this is really, this is it now.

Yeah.

What are we going to be able to do?

What version of Midjourney was that when you were?

I'm going to say three or four.

Yeah. Okay.

Going up to now eight now.

I think. I've lost count to be honest.

It was 2022, I think. So yeah. Okay. And that, yeah. A lot of fun. So, what else, you know, how, how have you progressed from creating artworks to building the things you've, you've built today? What, what, else followed?

I, I guess it was just kind of more of that. And, and when I look back over the last, the last few years, it's just continuing that spirit of play, learning a little bit more. And then once you, once I kind of developed enough trust or maybe more trust in myself than I knew what the hell I was even doing, you start to, I don't know, evolve and grow. And it's like you evolve from just chatting to an LLM one-on-one to then realising, oh, I don't know how to assign this thing to a task and, or, or an agent. And I guess you just get a little bit more systematic in your thinking, or how do I, how do I not just have singular conversation, but how do I kind of put these things to work in areas that are within an application? How do you support the, the, where I feel like it's really, really great. It's sort of those behind the scenes tasks that are just laborious and boring. And it's a bit of a churn, like going through and I don't know, analysing records and making sure that, you know, you remove all the false positives and just stuff that happens in any application. I guess I moved from playing, experimenting, having fun and slowly starting to creep its way into my real jobs, my real work. Yeah.

Slowly, slowly taking over. Yeah. Oh yeah.

Yeah. How do we get, yeah.

And so you mentioned before you're in Barnes and Noble and you're, you're, you're reading programming books. So have you always been engineer minded or development minded with a flair for creativity or where on that kind of left and right do you sit? Do you think?

I don't know, but my answer would be if, if I had a screen recording of what's on my screen, probably over 70%, I'm staring at consoles and terminals and rainbow coloured code. And like, and I'm like, how did I get here? This is not like the trajectory that I was supposed to be on. I, yeah, I'm very, I get excited by the design and the product that, you know, is possible and the experience that's unlocked on the other side.

And you feel unlocked, right? Like you, do you feel it's not that you don't want to work with other people because a lot happens through collaboration, right? It's brilliant. But you, do you feel like you have these ideas, you can now realise them. Yeah, totally. And you can get them somewhere before you show other people. Right. Totally. And it's just an idea. Yeah. And the interface to doing that right now is just text on a screen. Yeah. That's where you spend most of your time.

Exactly. Yeah, absolutely. It, a lot of, you know, the ideas that I have kind of live in that prototype space and it's just getting and massaging an idea until it's at a point where it's shareable or I even know how to communicate what the idea is. It's polished enough that maybe it kind of makes sense.

So were you building prototypes beforehand without this technology? So you, you're comfortable enough in dev tools to, you know, spin something up and deploy it, all that kind of stuff.

Yes, but it's changed. I think in the old paradigm, pre, pre AI, AI, you sort of had to get a client's buy-in on the idea before you could even move into building the prototype stage. Like that was, it cost too much to even do that.

100%.

Yeah.

And the idea is the thing that you sell first. And then if that works, then you get budget to actually realise it. Yeah. Yeah.

And the amount of times that those ideas died before they were even given, you know, the first bit of life support.

Yeah.

Yeah. It's just crazy how many times the ideas shut down and I get it, it's expensive. But then there was right around that period where I was experimenting with, I don't know, coding with AI for the first time outside of just like the, you know, co-pilot where it kind of fixes your syntax and stuff. Yeah. But really using it as a build tool, it, it occurred to me like I can just, and this is during VML when I was sitting there in my job was just to put slide decks together. And I'll never forget. And I was like, wait a second, I can just stick a link on this slide deck now that they can click on and it can go run the thing that I'm trying to say you should do. Yep. And there was a noticeable step change in like how many of the ideas that I submitted that had a link on it or you could go and do the, use the thing, how many of those ideas got through the door versus the ones that did not. And then I was like, this is valuable.

That's powerful. People, people need to see things. You know, it's hard for the general to, to visualise an idea, but to, if you, if you create something and share, then they're much more likely to get buy-in, right? Totally. That's, that's really cool. So you started out just introducing things and it started to take over because it converted better.

Totally.

It was like a sales strategy.

Oh, it was. And I felt like a sort of magic trick or something like, cause like the, the folks at VML, you know, hats off to them. I always felt like I had the most like, what's the word? Good people. Yeah. Imposter syndrome. Cause I was just like this digital designer developer, but then in this sort of creative space next level, everyone that's coming out of that agency.

Yeah.

And I was like, I learned that I could lean back on some of that development background and put some links in there and maybe get a couple ideas through where I was always felt like I was putting ideas up against these, these folks that were just brilliant at putting the best pitch deck together possible, which wasn't necessarily my, my sweet spot.

Creativity is hard. It's so hard. And I got a lot of respect for people that can do that. Well, I'm, I'm probably similar to you in that I think where my strength is, is I can, I can be creative, but not at that level, but I can realise things. So if I get a slither of something that I'm like, that's, I can, I can build that. I know how to make that a thing. Yeah. And then from there you iterate. And so it's, it's funny, the point you raise around it's, it's the cost collapsing, right? So this is a transition that I, I don't know how agencies are going to navigate this, but you're right. It's much more cost effective to just make something now and share it with them. But if you don't have budget to do that, how do you do that? So they have to look at their, their, their cost structure. You know, how do you, you know, is it you, you have a continuous retainer to just generate ideas. So it's like the business of agency has to change around this technology because you, by the time you've, you've had not, you can have an idea and by the end of the day, you've got a really strong prototype, you know, and you could share that with the client. The next day and say, Hey, what do you think of this?

Yeah.

And if you, if you're going to do the other way around, we're going to create, you know, pitch check for a week. The cost of that is more than actually producing a thing. Totally. And, and I'm using this technology now to skip the deck altogether. I just make websites. So my proposals, contracts, everything is just a full website because it's what the technology is good at and it's, it's much easier to do. So.

I mean like the actual proposal itself and stuff is, is what, wow.

Complete website because it's just a lot. It's a, it's a lot easier for agents to understand, you know, web structure and, you know, components and, you know, all that kind of stuff rather than I got to design a slide deck in, you know, Google slides or whatever.

I get it. I get it. I'm always like, what program should I use for this?

And then, you know, using something like using browser use in any of these tools is awful. So the time is just crazy. So yeah, it's an interesting time, I think, especially for larger businesses where they're, I think you, you've made the right decision where you can just be really nimble and say, I'm just gonna do it like this, you know, and, and see how it goes. Okay, cool. So that's great. So what happened? What happens next? You've, you, you've, you've played around with a while. You, you have VML, you're creating more things in your proposals. What, what led to the next, the next transition?

There was a, a specific idea that I was really passionate about and, and it got through the gauntlet of creative director decisions and stuff. And, and I just came, became absorbed with the idea really and wanting to build out the product. And I think through that process, like something, something kind of shifted where I was like, like you just said, I'm just going to go and do this my way and, and start building up. Cause I'd done freelance before and I was like, you know, I could probably do this again. And I guess little by little, I just started saying yes to, you know, referrals and little projects here and there. And I guess build out a kind of a support network and then, kind of moved from full time to half time to a quarter time until I could make that, that move. And yeah, all, all along still just kind of experimenting with ideas, trying to figure out what my workflow might look like, how I can, I guess cost effectively produce actual shippable products and also just having fun and trying new tools. I feel like I'm preaching to the choir here.

You can carve out a whole day just to play with new tools. Yeah.

Man, tell me about it. That feels like it's a job in and of itself these days.

Yeah. I'm fine. I do find it sometimes hard to keep up, but I think if you niche down, right? Like there's so much, I'm not going to, I'm not going to pay attention to it all because I can't, but like, if you've, you know, I want to focus on this. There's enough going on bubbling away in that one little space that you can kind of keep track of it.

Yeah, man. I feel like that's you to a T like, I feel like you are keeping on top of it all. When I follow like your stuff, you share and stuff on LinkedIn. I'm like, that's a gym. I'm going to take that. I've downloaded pencil today.

Yeah. Pencil's lovely. Yeah. I mean, I, I, that's what I want to do as well as share this stuff, you know, like I don't, I got no, I don't want to hold on to anything. This stuff is moving so fast, you know? So I wrote an article the other day about my own harness. I built a harness for agents to build software and yeah, I just, I just basically put everything in there that I, that I created within the blog post and you can point your agent at it and it will more or less build it for you. And I thought, well, why not? Right? Because your things are moving so fast. We've already moved on to loops. So now I've, I've built a gentic loops into it. So you just keep going and just share. I just think it's the right way to go. So.

Even, even I read the, I, I, had a bit of a poke around your website last, last night and by the way, amazing website. Yeah. Well, where'd you find the time to go and build that in and amongst all the things, but.

Well, that was, that was rapid.

Quick. Well, it doesn't look like it.

Oh, thank you.

Yeah. But that article, like even that, like that is such a nugget man of a goal getting to that idea that like you could build out agents that could harness other agents. And yeah. Yeah. I guess work together and work through a process. Yeah. It feels like super forward thinking and it makes so much sense after reading as like, oh, the way you're thinking about approaching using your agent is like running a team.

Yeah. Yeah. It, well, they are. They're just like, they're just like people. And you know, it's like, if you gave a person a job, right? And if you hired someone, you, you probably want to give them a job description. Yeah. And then we used to do that with prompts. You are this, and this is. Yeah. How you think, you know, so you give them the job description and then you, you want to make sure that they're doing that job. So you go check in on them and say, how's it going? You know, I'm off the rails a little bit there. We want to focus on this. So you're guiding them through and you're giving them the tools that they wanted to, you know, use for their jobs. So, you know, they, they get access to web browsers and whatever. And so it's basically just the same lessons we've learned over, you know, our own, our own businesses applied to these things. And I always found that like, you know, I kept getting this feedback from people and they're using it as like, oh, they're not very reliable or they go off and they do this. And, you know, it's like, well, yes, because you haven't given them context and you haven't given them a set of rules. You're the boss. Yeah. Like, exactly. It's your fault. It's bad leadership. If you give them all the context for things and then you, you create a set system for a workflow that you want, you can even enforce it so that they can't, they physically have to stop and go, hey, what do you think for you to ingest as a human? And learn it, happy with that, not great before we move on. Right. So it's not a new theory, but it seems to work really well. So yeah, I'm just trying to imply that stuff and share as much as I can. So I'm interested to learn about you though. So you have, you have a lot of ideas. You've, you've created a lot of products. I see things that you've posted online and I'm super impressed by them. I love OZ FM.

Thanks, man.

This is so cool. And I want to talk about, is it musical drones?

Yeah. Yeah.

Let's Zingo. Look, see, looks really cool. So you've got some really cool ideas, right? There's like four products there of varying levels of, you know, the different industries, the different ideas. Have, are these all you, are you partnering with people? When did you, it's a, there's quite a portfolio of products there. Yeah.

Look-See is that one specifically is myself and one of my colleagues from VML, but everything else, with the exception of there's one that's called Ayuki. That's also another different colleague from also from VML. But yeah, the other things are just little hobby experiments and it gives, gives me space to, to play and evolve and learn. And yeah. Yeah.

So you, so you just come up with these ideas and you know, what would it take you to go from an idea to, sort of a rough cut of a product that you've, you put out there?

So musical drones, that one, let's start there.

Well, that was, that was, that was, What is it? What does it do? Cause it's quite a cool concept.

It's hilarious.

So just explain it a little bit. Yeah.

I'll go back to the very start. I was, so I'm 44. I was trying to motivate myself to run more exercise and getting to that point where there's a noticeable difference of when I am working out versus not. Running out just to track myself, just to see if I could, just to see if I could get AI to even make my own running out. And, and I, and I did, and it was super simple. And I would always listen to Spotify while I was running. And my, my, partner, she actually came up with this idea and she was like, well, what if you could make it so that if you stopped running the app, noticed the, the lack of movement or your speed slowing down. And then music stopped. And I was like, oh, that that's, that's it. That's genius. Like that is as simple as that. A running app where you hit go, you start running. If you stop the music stops. And when you start again, the music keeps going. And, and I just thought that idea was amazing. So I tried to see if I could get AI. I keep pointing over here. Like my AI lives over here on my laptop.

There's, there's, there's just, there's just, yeah.

And, and so I kept prompting to see how far I could, if I could get a workable prototype of that running and, and, worked it out. It worked. And then, I guess it right around that time. But I didn't do anything with it. I just kind of left it on the shelf, so to speak. And, maybe a couple of months went by and I read this news article about Spotify and what, the owner of Spotify was now spending all of our subscription money on. And to be honest, it just, it just pissed me off. Like I don't usually get passionate about stuff. Like I I'm pretty placid, pretty peaceful, whatever. But just the thought of like this dude spending almost a billion dollars on AI powered. Sorry for those that don't know, Mr. Spotify is now spending or has spent, I say our money. I don't know that for sure, but where'd he get his billion dollars from on, investing into, AI powered military drones. And yeah, again, yeah, so the idea just pissed me off and, and I was like, oh, I've got this running app and I know it's connected to Spotify. Cause you've got to hit the button to connect your Spotify account. And I was like, screw this. And then, something about that news story and drones and a running app that was connected to Spotify, something clicked. And I was like, I wonder how hard it is to run away from, like, if you were out on, I don't know, the battlefield, how hard would it be to outrun one of these crazy ass. Yeah. And then I, so I asked myself, like, could I simulate that within the running app to give it some kind of, I don't know, gamify it. And turns out you can, you can, you know, figure out the speed of, one of these drones and you can mark your location and, and kind of predict, how long it would take to get to you. Yeah. And, and then it's just like clicked, like, oh, there's a, there's a reason for this app to maybe exist now. Yeah. If you go download it, it doesn't even work anymore. It's a shit app. It was, I just submitted it more as a. Yeah. Yeah. More as a protest to say, Hey, this is happening than anything else. And then it, it doesn't work now because I had to apply for a larger, API, what's the word more users allowed onto the platform. So you have to apply to Spotify to, to make your case. And I just said, Hey, you know, this is, this is an app that, you know, is against what you guys are doing. And I addressed it to Mr. Eck himself and, lo and behold, the app was denied. They, they, weren't on board with the idea. Surprise, surprise. Yeah.

Oh, that's incredible though. So, I think is a partially, the thing that's impressed me about that is like, you've, you've drawn a parallel between you're currently working on something. You're playing around with it. It's just sitting there and you, you wouldn't be able to get to the point where, without the technology. You know, I don't know what you use cursor or whatever to build something.

Yeah. Yep.

And it's just sitting there, but then you've drawn parallel, you've seen this thing and you've connected the two. So is that creativity is something that you've always had, or did that your time at VML just really supercharged that?

Bingo.

Connecting this with that.

Bingo. Yeah. I learned a few tricks from the folks there that put those decks together. That they, they, yeah, it's about finding that cultural moment that people are, you know, recognizing. And seeing if there's a connection and then that kind of gave it a narrative, which otherwise it would just be a shitty version of Strava. Yeah.

Yeah. It's not something that people, people would latch onto in a viral way, right? It's, there's no, as you say, narrative to it.

Yeah. Yeah.

Okay. Well, that's super cool. I love that.

Thanks, man.

What about, what about, okay. So, OZ FM, let's just quickly touch. I want to touch on a couple of these and we'll just see where we go. But like, I've talked to you all day, it seems. OZ FM, tell me about that and tell me creatively how you're using this technology to sort of grow, grow this and build this and design this. Cause I'm really interested to hear how you are using this technology from a design sense, not just a build sense.

Yeah. Okay.

Even if you are, you know, there, there, there's so many tools out there. I have this theory that, or not theory actually, I've tested it. I don't think they're very good at design, like really good design, but.

That's what I wanted to ask.

Well, let's go there. Let's go there first. I, I, I've been using pretty much every tool that comes out. I was reasonably critical of Claude design. For a number of reasons, but like, I'm not a, it depends on what kind of design you're talking about. Right.

Mm-hmm.

So I, I like to think of myself as a product designer. You know, I've, I used to do branding a long time ago. I've done a whole bunch of different types of design, post design for gigs way back when album covers and stuff. But, you know, I used to do, yeah, yeah. Identity design. And then, you know, you know, basic websites all the way through to, you know, now apps, which are more user flow based, you know, and design in that sense is all about, you know, thinking about complex user needs across a pathway, you know, multiple pathways that a user could take and all the things that could go wrong. And how much do you put in front of a user and how much do you take out to make it a really seamless experience? Right. And then you have to take into consideration all of the, the guidelines from the app stores, like Apple and Google have all these Apple human interface guidelines, et cetera. And you really have to have a lot of experience in using these tools to know what works and what doesn't and know when to use a certain interaction when not to. Yep. And so my experience with all of these tools has been that they're awful. I've not used a single. Spot on. We'll touch on it.

There was one out there that was just amazing, but.

Well, I say that in the past, so I will touch on Fable, but yeah, by and large, they've been terrible and any attempt to build some sort of system around them still hasn't been great. So the, the best results I've had was shipping from, shifting from, you know, just an infinite canvas where you're just blindly working with these tools and saying, design me this, design me that. To the thing that's worked is context. So. I've shifted my workflow to be really heavy requirements first. So you spend, as you've said, a lot of time in text and you are elucidating ideas. You, you know, you getting the system to grill you on, you know, what should this be? What shouldn't it be? And oftentimes you riff enough where you might change your idea, but it doesn't matter. You've just got to get through it. And, and once you have a really detailed set of requirements, you feed that into an agent as context and you say, right, these are, there's all the functionality that it has. This is how it all needs to work. And part of that can be design. And I have used all these different tools like impeccable. I've used a bunch of different, you know, design skills and anyway, I, I don't think they're all that they're very good, but what does work is at a basic level, sort of a standard set of primitives. So like you give it, you know, a description of the style, you know, it might be bubbly fun, blah, blah, You give it the colours, you give it the fonts. You might give it some reference images to draw from, but then you give it the requirements. That's when it starts to get a lot better. So using tools like pencil, you, you can talk to your code base. So you basically tell your agent, look at this feature and create a flow that addresses all of the needs of the users in the, in, you know, the PIDs. And it does an okay job. It's not bad, not great. And there's still a lot of flow problems. You know, that's inconsistent. That would change that. There's no offline state. Too many, too much text and they way overcook it. Yeah. But I will say, and this was even before it got banned the first time around Fable five was another eye opening moment for me because it not only designed the flow in the way that I expected, it actually improved things. So it understood the requirements and it improved a particular flow on a product I was working on where I thought that's actually really smart. You know, it, it kind of was like, Whoa, I'm dealing with something else here. And now it's, it's been, it's just been turned back on and I'm just, I'm, I'm this poor thing. I'm abusing it, but it's, it, it's getting there. I w again, I wouldn't say it's, it replaces senior thought because you have to have everything in your head anyway. You need to understand what you're building. If you just hand this off and you go, you do it, you do it, you do it, you do it. You're going to get rubbish. So you, you have to collaborate and, and, that takes time, but I see a world where this technology gets very good because you're seeing the progression. Mm-hmm. And I think as designers, we have to acknowledge that and say, okay, well, we can't be too elitist, right? Like I saw this backlash for the OpenClaw app. I don't know if you, you follow OpenClaw. They're, they're an open, so there's still an open source project. They weren't given hundreds of millions of dollars by OpenAI. So they're still just supported by maintainers and they, they released a mobile app. And I mean, it wasn't very good. The design looked really sloppy and they just, you know, obviously it looked fire coded basically. And designers on, on social platforms just raked it over the goals.

Yeah.

And I thought it was really sad because, you know, firstly it's open source. So if you, if you don't like it, design something better, but don't, don't be, don't put yourselves in a position as a, as a, a, group, a culture, like as designers, where you, you, become, you have this elitist viewpoint.

Mm-hmm.

You should try to help people because this technology is only going to get better. And there will come a point where that mid tier of designer just gets wiped out. Well, unfortunately I'm seeing, I'm seeing the trajectory. But if you're using it to get to that higher tier where you're doing more of the thinking and you know what good is, you're going to be fine. And people are still too busy to do it themselves. So anyway, that was just a little bit of a side, but I think that, yeah, Fable was great. Yeah.

I still, I haven't tried it myself yet, but I've seen like the stuff on social media and it just, yeah.

It's good.

It's, it's, holy cow. Like is, I can't believe it's gotten this far. But I'm always like a little sceptical. Like some of the stuff you see shared, like, okay, how much of that really was made by Fable versus did you have to coach it along, et cetera?

Yeah. But I think you're right. As a deep, deep, user of it, you know, I see all the posts where, you know, it's some hype video and, and most of it's nonsense. Mm-hmm. I'd say.

Okay.

Don't have the time to reply to them all, but yeah. It is pretty impressive.

I've seen the, like the examples and stuff where they say that they prompted to make like a GTA six or something. And it's just like.

Yeah. It still takes a lot of work, you know, as you're saying before, it's a different kind of work, but it's still a lot of work. It's still a lot of craft involved, which I like.

Yeah, totally.

But anyway, what about you with, yeah, your, your approach to using it for design and, and, OZ FM.

I guess, first of all, the design on OZ FM is terrible, but I, it was, it's just sort of a, a hobby project, just to see if I could pull it off. The, the original idea like started as, actually back at VML. There was a, I can't remember the name of the band. One sec, Chiquitita, the, the, not the Bee Gees. Damn it.

It's just, stuck in your head.

Yeah. Now I can't, I'll never forget it.

It'll come to you if you keep talking.

Yeah. Anyway, so there's this project. They had this, this, huge famous song, Chiquitita and they way back when in the seventies, ABBA, that's who it is. ABBA. Yeah. They, they had put it out in the seventies and they gave the royalties away and they said, and they gave it to UNICEF a little bit of. Yeah. Love and history. Yeah. Abba history. And UNICEF actually came to VML and said, Hey, the biggest donation that we ever received from anyone was that song because the royalties from that song. It was just came in at, in the millions.

Whoa.

And they were like, can we do something like that again? How could we strategically encourage more artists to do the same? You know, is there a Taylor Swift out there that has a, you know, not a massive hit, but you know, a mediocre one that she could give away and. You know, help a lot of people in need. And so that was the brief. How could we encourage more artists to do the same? And I had been experimenting with seeing if I could get AI and voice AI to run a fictitious radio station. Yeah. Just, just for fun. And, and I'd connected it also to Spotify and I just wanted to see if I could rebuild that sort of seventies, California, you know, beach radio vibe. And if I could get a DJ to come on and talk about, I don't know, relevant news topics, even introduce the next song, know the last one that was played, take live callers or chats or whatever. And so there was that idea and I was like, Abba, maybe there's something here where we could promote artists that are giving away their music and stream it through that. And so they love the idea. Cool. Let's, let's, do it. But then the war in Ukraine, Ukraine broke out and UNICEF's focus kind of shifted a bit, blah, blah, But then that idea, I was like, well, I'll still go and build it and to still make a AI powered radio station for fun. And so it changed names a couple of times and it just kept working on it. And, and then I got to the point where I was like, you know what, I'm just going to release it and put it out there and it's free and it runs off of sound cloud and you can play with it and turn DJs on. And yeah. Very cool.

Yeah. And so what is your, what is your process for building something like that? Is there any sort of tips you can share for viewers where you go from an idea to a product, you know, as a designer? Yeah. What, what, how do you go about it? I actually do because it's a cool idea and it's, and the app, the output's funky too. Like it's actually quite, you say it's not a good designer. I disagree. I think it's fun. So how do you, what's your approach to that sort of stuff?

So it's just start simple. This is going to sound so basic, but just start talking to AI about it. Even like, you know, here's kind of the rough shape of the idea here, the features that I wanted to have. It was very similar to what you were just describing and establishing that PRD. Here's the requirements. And you kind of start there. And although I don't think I was doing it back then quite like that, I hadn't quite evolved to even that level of thinking. Like start with that plan document and that becomes the, you know, the whole, the Bible, the source of truth.

Yeah.

You know, the defining thing. It was a bit more like, can we make this feature? Can you make a connection to this platform connecting through this API? Here's the documentation for SoundCloud's API. Okay, cool. Now we have a connection. Next thing. How do I get a button on the screen that you can hit next? And it goes to the next thing. What is a good delivery mechanism for, yeah. Figuring out when a DJ should come on and how does it know the context of what was last played. Just talking to it about everything. Yeah. And kind of following its suggestions and then leaning on, I guess, my prior experience a little bit and saying, no, maybe we should do it this way.

Yeah. That's good. But you have a, you have a level of sort of professional experience in these types of things anyway, right? So it's like you, you're, you're brute forcing it, but you're brute forcing it with like a lot of knowledge already. So as you say, I've got it, I mean, this, you know, a little bit, yeah. The, the average, the average person, you know, I speak to a lot of designers where they're not from this world, you know, they're not, they're not built. And, and I keep, I keep making the point that designers need to be builders, not developers, but builders, because you need to actually see the thing. You have to create the thing to see all the edge cases and stuff. And sometimes they get blank faces where they're like, well, I don't want to do that. Like designing in my, I don't, I'm not a developer. Well, I'm trying to figure out ways to, to help people understand. It's, it's, quite straightforward. You just have to, as you say, you just talk to it. And yeah, I, I, actually talk, you know, I use whisper flow and, and I just talk to my machine and my wife thinks it's really weird. But if you, and you, if you saw some of my prompts, you think, oh God, if you don't know, it's the perfect technology to just feel stupid. You just, you just talk through it. If you don't know something, you just go down these rabbit holes. Totally. And so that is, is that been your progression in the last like six to eight months or so? Like just constantly.

For sure. Like, I, I think, yeah, it was, I think it was just born out of that spirit of play and never trying to get too serious about it. And like, like you said, not, not being fearful of, you know, saying something dumb to it. And I've just learned that you can just bring your natural language to it. And it's amazing at figuring out and inferring what you meant. And the more, the more that you give it, the better. It's not, it's not like you need to define everything perfectly or, you know, you see like a lot of stuff on social media where people are trying to slog their perfect prompts that are like, it feels like they're trying to sell some system. That's like, this is perfectly written and has all the definitions needed. Like, that's a thing. Like there's a syntax or something to AI. And that's, that's not, that's not the case.

Couldn't agree more. And then it creates this. There's no such thing as a perfect prompt.

Yeah. And it creates this perception that like, you've got to be, yeah, knowledgeable in all the prompting and prompt engineering. And even that, like, I get it. There is, there is totally a role for prompt engineering and there's some do's and don'ts, but it just makes it sound harder than it is.

Yeah. Yeah. I think you're right. It's, it's not as hard as, as, people think it is. And on the prompt thing, it's like, if you, if you wanted an app and you hired an engineer and you basically just sat down with it for five, two minutes and said, right, I want an app that does this, this, and this, and it's got to do this, this, and this. And then you said, off you go. And they disappear for, I don't know, a month or two months. They come back and here's your app and you go, oh, it's awful. Yeah. It's all I wanted. That's not how it works. You have to collaborate and work with, right? So I think you're right there. Yeah. I have a bit of a, a side question for you. I'm, I'm interested in your, I'm interested in your, your approach to this or what your thoughts are with this anti AI sort of sentiment that I see within certain communities, especially with regards to creativity and, and just what your thoughts are on it. Because I, I certainly empathize with a lot of the diffract, the diffractors and the, and the issues, but I also see that the, you can't turn the switch off. And I do think that some of the concerns are overblown and I see it as a, as a, as a tool in more of an additive tool to be more creative. So from, you know, you and you and I think are quite similar. So I was very interested to hear what your, your sort of thoughts are on that. Is it doom and gloom? And, and, what would you say to the concept of, you know, this stuff, these companies sort of scraping everyone's data and spitting it, spitting it back out iterations of it.

I, I, I'd like to think that I'm generally a positive person. So I try to, to not go to the doom and gloom. It's definitely like always been a thought and in the back of my mind, like, what does it mean if this becomes easier? I've built my life around learning this, the skill, what does that mean? But I don't think it's doom and gloom. Like if I look at my own just personal experience over the last couple of years with it, it's been absolutely positive. I, I, hear the concerns and, and, there's a big part of me that totally resonates with the anti AI movement. Don't get me wrong. Like I, I, and I'm starting to ask myself, like, when is the right and wrong time to use it? For example, even on OZ FM, the very, one of the very first iterations on it, I had cloned David Bowie's voice, just out of, out of fun. And then I sat with that for a while and I was like, you know what? I don't think that's the right thing to do. And so I've shifted away from that. And, but there's also the, I guess the environmental concerns and how much power we're consuming to, to produce stuff. Like I think about some of my prompts that I let through and I'm embarrassed to admit this, but like, you know, there, there are definitely some tasks that I could do that. I'm giving off to AI at a micro level that like, come on, man. Like you don't need to be doing that, but you just sort of do, and it becomes part of your thinking process. And, but yeah, it's, it's, interesting. You ask that. Cause that was actually one of the questions I wanted to bring up with you was like, what have you been noticing this sort of anti AI sentiment?

Yeah. I, I really have. Thanks for sharing that by the way. It's good to get your feedback. I, I'm definitely seeing there is, there's people that just don't want to bar it at all. And they'll like change careers.

Yeah.

You know, they don't want to do it and fair enough, you know, you're free to do whatever you want. I think some of the energy sort of concerns are a little bit overblown. I have a different opinion on energy usage. I think that energy usage in itself is not bad and that the quality of life goes up if you, if a society is using more energy, which means that the focus should be on energy generation, clean energy generation and efficient energy usage.

Yeah.

If you're going to say, oh, energy usage is bad. Well, where do you stop? Where do you draw the line? Like you want to go, wear leaves and live in the forest? Well, I'm sorry.

Yeah.

Energy consumption is a part of modern society. So it's where do you, where do you use it? How do you use it? The water usage, you know, most people get that wrong in the sense that I think that these data centres just consume so much water, but it's a closed loop system. Most of them. So once they get it once, they more or less, they don't use as much. Right. And so it's the, you see these stories. They get perpetuated in media. And I think that there's this anti sentiment. This comes in and people just start to passively absorb, sort of read it and take it on board. So I think that's less of a concern for me. And I do think these things would be a lot more efficient. Right. So the average, the human brain uses like, I think it's like 20 Watts or something consistently compared to the power usage of these models. Right. So you think about what we're able to do with our little meat computer compared to these tools. And you see, there's a lot of optimisation, right? Where this is the worst it's ever going to be.

Yeah.

And we're already seeing that with like Chinese models. They're, they're optimising these things. They're getting them faster and faster and faster. So it'll get a lot better. It'll become more ingrained and less of a concern on, on, that side of things. So that, that part, I'm not too worried about the, the creative, you know, copywriting sort of concern. That's, that's real.

Yeah.

I, they, they, I mean, these companies have just scraped the web basically with no attribution. And, you know, I, you just generate, you know, it was when OpenAI had their image generation tool and everyone said to make me a studio Ghibli image of this.

Yeah. How OpenAI kind of endorsed that and pushed it.

It just took off. Yeah. And, and then they made it harder to do. Right. And then the models get so nerfed because you can't do anything with it. Cause it triggers all these safeguards.

Yep.

That it's, there's a fine line there. So I don't have the solution to that. I, I don't like it. I don't have the same model or some way that people, you know, it's like, if it's like streaming, right? If you, it'd be great if you streamed audio that everyone that contributed that got, you know, continuous streamed money. But, but the system has to change. It's a different conversation. The entire system needs to change to support that sort of thing. So yeah, that is a concern for me. And I, I try not to, and to your point where you, this is, I really like what you said, where you said, I, I experimented with it, but I'm not going to steal David I always voice. I felt bad. Yeah. And that's the right way to be. I think to make a conscious decision to say, you know what? I'm not going to do that. And I think that's the way that we should approach it until there is a better solution. But I don't think just shutting down and just not using it is. Yeah.

It's used that phrase. It's throwing the baby out with the bathwater. Like when you, when you think about the, the positives that's come from it, like, I'm sure you can appreciate this. Like pre AI, there was so much work that you did not communicate all of the depth of it to a client that you were, you know, you would slave over doing stuff. And that's not fair on the developers or the designers really. That's not like a, to sit in front of a, in my opinion, to sit in front of a computer screen day in and day out building these digital tools. That's not a valuable use of human life really. Yeah. And to have AI kind of step in and then take some of that, that, stuff out, I think it's great for, for people and yeah, people's livelihoods that outside of work, like the, the amount of time that it sort of gives you back, I think is absolutely worth the, you know, the, the, stuff that we're going through to get it there.

Yeah, some people, some people derive, there is some people derive great pleasure in, in the, in, in, you know, if you're talking about development, right? They're actually writing the code, you know, and they've spent their whole career doing They're doing that and that's what they want to do. And that's a really challenging position to be in because that role is dramatically changing. We, they now have to become orchestrators. They have to understand a system to know where the risk is in the system to make sure the agent doesn't screw it up because the economics don't work anymore. You know, a client's not going to pay you the, the, rate of the, the old way. It just doesn't make any sense anymore. So what can you do with the time you get back? And that for me is the really exciting part. And that's about, I think it's about being more creative and starting podcasts and

Doing the things that you always wanted to do. Really?

Yeah. Yeah. We've got more time. So I've got, I've got agents writing PRDs right now.

I, I saw the coolest job title on that note about engineers the other day, and this guy had it on his LinkedIn and he put vibe coder audit auditor. So like five coded apps auditor. And I was like, that's it.

He's got plenty of work.

Yeah. You always want that person, you know, looking over your shoulder. That's got the experience that can say, Oh, I don't know about that one.

Yeah. Let me help. Yeah. Open, OpenClaw could have used that. So is there, is there anything else you wanted to convey any, anything you wanted to share before we finish up?

I mean, yeah, not really. Just don't be afraid of it. You know, experiment. Yeah. Play and see where you get some of those ideas that you've always had in the back of your head that you never thought would see the light of the day. Give them a shot. And you know, it might, you might be surprised. Yep. Yep.

All right. Well, great. Look, I'll tell you what, it's been an awesome conversation. Thanks so much for your time. I really appreciate it. We should, we should catch up outside of this. Absolutely. Talk for hours.

Yeah. Same, man. Cool. And enjoy your weekend.

Thanks too. You too. All right. See ya.