How Creators Are Using AI Agents to Work Smarter | Mike Russell #664

Welcome to The New Media Show with human host and Podcast Hall of Fame, Rob Greenlee All right, well, welcome to The New Media Show. This is episode 664 and I'm really excited to be here again with you doing the show. Today, the topic is going to be building your AI employee and how podcasters and creators are running smarter with AI agents these days. I know this is a very cutting edge topic, but I've got Mike Russell with me who is the founder of creatormagic.ai and has been a long time audio producer with his company Music Radio Creative, which has been helping podcasters for many years create their lead ins and bumpers and all sorts of stuff. So he's heavily into the audio side too, but yet he's really been pushing hard into the AI side and video, which he and I are kindred spirits here on that kind of stuff. But Mike, thank you for joining me on The New Media Show. It's great to have you on. Rob, man, it's so cool to be back on the show and I'm excited because a lot has changed since we've last spoken. Yeah, I've seen you a few times at conferences here in North America. I'm not sure that we've caught up in Europe at any point, but I've seen you once or twice in London. I've seen you there on the floor. Yeah, I've only been there once. It probably happened back in 2022, but let's go ahead and jump into it, Mike, and talk about what you're doing here. But most podcasters think about AI as a tool at this point, something you can use to clean up your audio, write your show notes, and create a transcript and things like that. But Mike, you've been thinking a little bit bigger about this topic and trying to build AI into kind of your own personal employee of sorts. It's almost like a team member over here that's doing work for you. And your agents are doing what for you at this point? Just curious. Totally. It's a great question. So yeah, I'd like to maybe take two angles on this. So first of all, the idea of building your own AI employee. So if you, it's the same working with AI as working with humans. If you know what you want, you're going to get a good result. So then you can start building that employee to do all those things that maybe you're not so good at in content creation. That's the first angle. So you can get, say you're really good at recording an audio podcast. You can get it to help you with the video and the thumbnails and things like that. But the second angle I'd also really like to get in is if you are still using AI to prep your show, write potential questions for guests, create scripts, and maybe generate an image or two as a thumbnail or whatever, then you're leaving a lot of leverage on the table. And actually, this camera angle that you're watching me on now on this show has been dialed in by AI. So again, I've been using AI to control the whole of my physical studio. So yes, you can use tools like OpenClaw, Hermes Agent, and I'd like to talk more about them throughout the show. But you can also plug these agents and things like Claude Code and Codex from OpenAI into the things on your home network. So right now I'm lit up by two Elgato key lights. I've also got Philips Hue lights behind me lighting up my display, which obviously needs to look good as a video-based content creator. I've got a Sony camera that is 4K that I'm talking into right now. And I didn't know this, but Claude Code told me it has a way to control all the settings on it via your computer. So I said to Claude Code, I didn't even worry about finding out where I needed to get the tool. I said, you find and download the tool, make it work with my camera, and start dialing settings. And it literally changed me from being a flat, non-color-graded image to a cinematic look by controlling my lights all around me, controlling the shutter speed, teaching me about zebra lines or zebra lines, if you like, on cameras, and so much more. So I would encourage you to dive in at the deep end and use AI for everything, and a lot more than just text and image generation, because it can also start controlling the real world. And I think as we go further in the episode, we might talk about how we'll be interfacing with AI in the future too. Yeah, definitely. And it's good to see that we have a common interest in our color backgrounds with our backlighting. Yeah, I built this all out back during the pandemic when everybody was shifting towards getting video going because everybody was doing Zoom calls and all this kind of stuff. So that's when I put in the effort with the studio lights and the higher quality cameras. And I've got a similar thing. I've got a 1080p 4K camcorder, and then I have another 4K camera here. So I can do 4K or 1080p. So you've got, basically, if you look at what's going on with YouTube, YouTube prefers 4K footage. That's the impression that I've gotten. But if you look at if you're doing podcasts still where you're pushing your episode out in RSS or you're planning on working with HLS streaming into Apple Podcasts, most of the platforms are really only at the stage of 720 right now. And that's, I have mixed feelings about that discrepancy. So it's almost like we're going to have to create two different video versions of our show to tap into different parts of the infrastructure. And so I think that this is probably a little bit off topic than talking about AI. But I think the bigger message that I'm trying to say here is the complexity of being a creator now. And it's not so much about being a podcaster anymore. It's about being a creator is putting us in a situation where what you're talking about right now really taps into the core needs of the direction that we're all heading. And that's complexity. And how can we use AI to cut down on that complexity to make us more efficient and keep our costs down to some degree? It's expensive to hire people to do various parts of our production and our post-production and things like that. And the publishing part too. So these AI engines that you're working with can make us much more efficient. But it's one of those technical hurdles that I think people are going to have a hard time getting over. And it's going to take time. And maybe these software companies, these AI platforms are going to have to build in more simplistic ways of utilizing them. And I think they've started to do that. If you can talk about that part of it, of this transition because of complexity and how AI can help with that process. It's a little more than a tool now, isn't it? Definitely. I think that's a great statement as well. Because if you think about where the moat is now as a content creator, I think it lies firmly in how good the content is you're producing looks and sounds. So we were told maybe two or three years ago, if you were doing an audio podcast, it's probably a good idea to at least get a webcam in the room that you're creating that content. Right. So then you can throw it on YouTube. You know, Spotify is supporting video podcasts as well. Like you say, Apple also taking that on. So it was going the video way. But now more and more, particularly I think late last year when the CEO of YouTube said YouTube is the new TV. That was his exact statement. It's like, oh, this is no longer like 720p, 1080p. We have to be 4K. And actually, Rob, just before I got on this show with you today, I was looking at my YouTube stats thinking, is that really true? I, you know, I still see comments from people saying that your screen shares are a bit too small for a mobile phone. So I'm like, yeah, most of my viewers are on mobile. And I go into YouTube analytics for the last 90 days. One third of my audience and the biggest slice are watching me on TV. And the, I think it was something like YouTube is now the number one app in terms of watch time in the United States, hands down. So it's like, this is where we need to be. And we need to be there in 4K. And the new moat is creating content that looks good. You know, has that cinematic blurry background. Awfully lit if that's the look you want to go for as well. But you have to be more than just like, I'm going to throw a webcam up and do this now. Now it's like the name of the show. It is your, it's new media. It's the new media. This is what it is. Yeah. I think we have moved from the, you know, the podcaster, podcaster creator as being really what that new media creator has been for many years, for 20 years here. The podcast industry and then the growth of YouTube that has been parallel have really been the catalyst to create this quote unquote new media, right? Which is really a digital replacement for broadcast. And so that's kind of what is going on here. And a lot of people don't like to accept that's been going on, but it's been going on since the early days of podcasting. When all the broadcast companies were doing video podcasts in the early years, the only reason they got out of it in the early years or moved over to YouTube was because they couldn't afford the bandwidth or they didn't want to pay for the bandwidth because they hadn't figured out how to monetize it yet. Because commercial broadcasters are commercial, right? They need to make money because they have shareholders and all this kind of stuff. And that's what caused a lot of them to stop video podcasting in the early years of podcasting was because there was no, you know, monetization technology layer that was layered on top of podcasting in those early years. And so now we've kind of gone full circle. So we went through this era of audio only, which basically started in, I would say probably 2011, 2012 is when we started to shift to seeing podcasting as a audio only medium. And that's, that really got reinforced with serial in 2014. And as we, you know, moved up into, I would say it started to shift back to video back in 2020, 2021. And, you know, so we saw this big shift back. And so, you know, it's interesting how those trend lines have moved through the years in this medium and how people's perception of it hasn't always aligned. So you're one of those rare guys that really came out of audio and really went all in on the video side, kind of like me. I mean, I started on the radio back in 99 and I've spent most of my career on the audio side, but I've always had a fascination with video and worked with it a lot when I worked at Microsoft. And so it's part of my native orientation to this, this medium is to think that podcasting in the earlier years and in the present tense is audio and video. And it's really always been that. What's your thought on that transition that you made going from really being an audio focused creator and a service provider to this video and now AI approach that you're taking? Yeah. Yeah. Well, it's a good question because I'd been putting the basic videos out on my YouTube channel back in 2010, I think was my first video. And those were video tutorials. And I remember for the first two or three years, and you'll still see this if you type my name into YouTube and look at my first videos from like 16 years ago, I did not show my face on that channel for like three or four years. And then I went to a conference. I feel like it's a conference you might've gone to Rob called new media expo. Did you go to new media expo? Oh, I went to all of them. Yeah. There you go. So I went to a new media expo and I remember when I was there, I had this distinct conversation with Ray Ortega actually. And he said to me, he said to me, you know what? Those videos are cool, but you should put your face on the camera. I was like, oh. That idea came from Ray Ortega? Ray Ortega. Yeah. It was another audio guy. And I was like, there's no way. There's no way. He just, he looked at a few and he said, yeah, they're good. They're nice tutorials, but your face on the camera, Mike. And I was like, no, I like you, Rob, I worked in radio. I'm like, I've got a face for radio. There's no way I'm not doing this kind of thing. And yeah, you can wind back through the archive and you can see how embarrassing it was when I was sitting there in my, in my attic working away, filming at a weird camera angle. But the one thing I would say is, yeah, the, the, the bar keeps raising to become a content creator and become a content creator that can get eyeballs. Right. Right. So this is really important. And a lot of people are not comfortable going on camera. I think that's a common theme even today with creators. Some of us may not even be comfortable talking into a microphone. Like I hate the sound of our own voice, but at least no one's watching you and you can do it in your PJs. So you can start there. But now people are like, no, no, it's got to be video. And no, no. If you've got wrinkles, like I have tons of them in my forehead, you see them all in 4k now, which is crazy. But you've got to do it. So you've got to start somewhere and you've got to put that content out because the more you do it, the better you will get. And it is a journey. It's like life is about the journey, not the destination. Right. So if you literally wind through my archive or anyone else, even watch Mr. Beast, Jimmy Donaldson, watch his first videos when he was 18 years old. Like they are filmed on the jankiest camera. Like actually, Mr. Beast, if you go through his archives, he actually did a tutorial about audacity. And I'm like, I'm wondering, did that dude watch my channel? Where do you get those ideas from? Right. Yeah. I'd like to think anyway, but everyone has to start somewhere. That's my point. So why do you think that it's so important to do video now? I mean, I have my opinion about it because I've seen the path over 20 years, but why do you think it's is now and why we're shifting over to a video first kind of a mentality for the creator space? Oh, that's something I need to give some really deep. Because I was going to, I was, my initial response would be to say it's more personal, but no, it's not because I think audio is more personal. Well, it's got a different, it's got a different dynamic to it. It's more, more imagination. It's less, less focus is needed for audio. I think then video, you know, it requires hearing as well as seeing where audio just requires hearing. So it's part of a, you know, I think that there will always be more value in just the flexibility of consumption on the audio side. And I don't think it, I'm really concerned too much about audio, but I guess one thing that does jump out at me that makes me worried about audio. And I think you might agree with me is the fact that the industry is shifting more towards getting the audio from the video and making that the new experience. That has me a little bit worried because I don't know about you, but historically my work with video has required a different edit from the audio. Right. And to create a better audio experience, it's kind of hard to have it be the same thing. Right. There are some creators, I think that do incredibly well, like Stephen Bartlett from a diary of a scene. But I think there needs to be some retraining that happens to creators to understand, you know, walk in that line. Absolutely. Because you're right. A lot of people make the audio audience, the poor audience, right? So they'll say, oh, we're showing you this now and doing this and that and that. You know, someone like Stephen Bartlett will say for those listening right now on the audio version, my guest is holding up this da da da da da and it looks like this and blah, blah, blah kind of thing. Still makes you feel as an audio listener. I want to watch the video now, but you know. Yeah, that is the thing. And yeah, how can you how can you do that? You know, because ultimately video is visual and it commands attention as well. Right. Whereas audio can be passive. You know, you can listen while doing other things right with the video. You really have to be locked into that video at the time. And therefore, it's important when you're doing video that you're doing something that has retention, that retains people because otherwise they'll switch off. Whereas someone, you know, that's why they put a noise the heck out of me. But they're now starting to put these like mid rolls in the middle of podcasts all over the place. So like I'm doing something else. I'm like doing some exercise or washing the dishes. And I'm like, I've got to figure out the shortcut for my AirPods to like skip forward 30 seconds. There must be a magic tap to do that. I don't know. Well, ads are supposed to be annoying, right? That's the premise of them. Yeah. So but yeah, that's always been the friction point in podcasting too or any kind of content creation is how do you monetize them? And if you put ads in there, there's been tech that has enabled people to skip through it. But there's also if you look at cable television with the streaming platforms, they've disabled that ability to skip too. So you've got all sorts of tension going on there with that. But yeah. And it's annoying how targeted they are now because they as you download the moment you do, you know this, when you download, they check your IP. They probably get a load of other data points. Oh, yeah. And they're like, ah, this guy likes this and that. I'm going to send him that advert. You know, it's wow. It's freaky. Well, and that begs another conversation too is just the metrics of all this stuff too. You know, as Apple rolls out HLS streaming into the Apple podcast app and things like that for video, there's going to be able to be deeper metrics. And so we may see, you know, I think that there's been for a couple of years now, a lot of metrics envy about YouTube. As YouTube is seen as really the, I think I saw in a newsletter today, the app YouTube's studio platform is like, it's called the goat. I'm sorry. It's the greatest of all time kind of analytics and. Tell me a little bit more, Rob. Tell me a little bit more about HLS and what it will mean, because I'm not familiar with what that will actually mean for the future of streaming and video. Like, what do you mean that, that it will give you deeper analytics? I'm really curious. Well, I mean, if you think about how video was made available through RSS in the past, it was just like the MP3 file was just a download, right? So you didn't really have a lot of play deep play metrics on it. The delivery of that file was, was hard to track in any kind of great detail. You could get a ballpark on it, but it was not very precise on how far people consumed before they dropped out. And especially if they downloaded it to, to the device and played it later, there wasn't a lot of metrics on that. So this HLS streaming, which is going to, that's happening is going to give the opportunity to have the same level of metrics that YouTube has, because that's the same technology that YouTube uses to deliver the video formats to all their devices. It really gives an opportunity for podcasting to kind of catch up to the metrics that are, that a lot of the ad buyers are looking for. How long it's going to take for us to meter all that stuff at the same level that YouTube has is probably years off, but I think it's coming. You know, I think the ad buyers are expecting it now, even from podcast consumption side now. So, and podcasting now will be able to directly compete with the consumption behavior of platforms like YouTube. So, so, and it's going to enable the possibility of live streaming into the podcasting space, which there's been tags. It's called the lit tag as part of the podcasting 2.0 initiative that will send a ping to a podcast app to say, this show is now live and you can click a button and consume it live. So that kind of capability is possible. It's not widely deployed yet, but all these things are kind of moving in a direction that hopefully will catch us up to, you know, what YouTube has been doing for years now as, as well as the other social media platforms. So podcasting has been falling behind, I think in a lot of ways. And this is why Apple made this move is because they want to remain relevant in the new media landscape. And it's a little bit, um, you know, that is a leading question to the whole thing about why video now. My opinion on that is that video has always been important to, to humans, right? So, and we just chose as an industry, I mean, the podcasting industry that we were going to focus on audio. The commercial radio people came into the industry and really had a vested interest in converting this medium away from video. And YouTube had an interest in that too. They wanted to get all those creators onto their platform. And so it was like this big sucking sound out of the video podcast space in the early years into YouTube or people just stopped doing video for some period of time. Yeah. I'm talking about the big major media companies and they just stuck to their lane of broadcast. And increasingly they were working with YouTube, but there was a lot of legal lawsuits and things like that going on in the early years around copyright on making any of that content available on YouTube. But all that got sorted out. And so now we're on this path video now. And I think it's also a generational shift to the younger generations have been more, you know, involved in video like TikTok and shorts and YouTube. And that growth has really propelled us into this new video era. So let's change gears really quick and talk about your community that you've built and what's going on there and what you're trying to do with your platform here. The creator magic, I guess you've got like 12,000 members of this community. How are you helping people in this community? Is it helping them adopt AI technology as being a creator? Yeah, it's interesting actually how it's evolved. So the new YouTube channel I started just over two years ago now, actually just a couple of weeks ago, hit 200,000 subs. And it started out very much as a channel to just help anyone who wants to play with AI and understand how it works. And it does in a sense, I introduced different tools here and there, especially like talking to content creators. And then over time, I started to play more and more with vibe coding tools like Cursor, like Claude Code, like OpenAI Codex. And I've suddenly found that my target, the person that I seem to be reaching the most now with where the channels evolved to over the two years, are 20 plus year senior developers who got like completely faded out from coding, like hated it. And then all these new tools came along and they've suddenly had their spark reignited by watching a creator magic video. And that makes my day and blows me away, actually, because I'm not a developer of any kind of QDOT. You know, I've copied and pasted before AI existed from Stack Overflow, but that's about the extent of it. So to be actually playing with tools and have them write most of the code for me in the things that I teach on the channel and also in the way that I create tools for my own content creation is really quite stunning. And I think it just shows that you can do so much with these tools to make your life easy. And I encourage everyone to have a go at it. So I think that's the premise of the channel and where it's grown to over the two years. I love it. So I've never thought of you as a kind of like a very highly technical geek would venture into so deeply into this AI technology with what you're doing. I mean, you've been so human focused and audio focused and stuff. This was a big shift for you, I think. Yeah, it really is. Because every so often I'll get a comment. Oh, it's the Adobe Audition guy. It's doing that. Oh, yeah. I've been at events and conferences where you've been on stage. I've been on stage talking and I've been doing that too. And I think we've done panels together, I think. That's right. Yeah. But yeah, I've always known you as like highly, highly produced audio producer. Right. And so this shift towards video and AI, but you've always had your finger on the pulse of where the industry and where creators are. And I think that's the reason that you got here. And I think I'm very similar to that too. And, you know, Todd Cochran, my former co-host of the show argued a lot about this shift towards video. He got tired of hearing me talk about it. And then it evolved into talking about AI. And he was a little bit two-sided on that AI topic. And I think a lot of people are about this. They'll say something publicly about it. That's maybe a little bit, eh, it's not all that important. And it's, you know, it's not very good and all this kind of stuff. But I knew behind the scenes he was working with it a lot. Absolutely. You know, so I think there's a lot of people that are living in that realm because they want to give a perception externally that they're human-centric and human-focused. Because there's this perception that AI is going to threaten humans, you know, livelihood and existence and stuff. So there's a lot of fear going on around AI. I would dial into that. I would absolutely dial into that because I would say with absolute confidence right now at the moment that this episode is going out, AI lacks something that we still have as humans. And that, in a simple word, is taste. And it's true. It's so true. You can generate the largest amount of AI slop if you want. I could sit here tonight and set up 100 automations that could pump slop out onto the internet without thinking about it. And I've done these experiments over the years. And it's still not there. So, my goodness me, when 11 Labs was becoming a big thing as the best text-to-speech and now there are a bunch of others that are doing it as well, really well. But 11 Labs came out and I was like, this is actually really freaking good. So I started, like, cloning my voice. And I was like, right, I'm going to pump out podcasts that will just do this, that, and that, and stuff. And, you know, I see some people on YouTube. They're like, there's one guy and he's, like, publishing a video an hour about, like, trends. So he's clearly got some, like, Google Trends scraper plugged into 11 Labs, plugged into something else. And he's pumping, like, probably using a Haygen. And he's, like, pumping a video an hour about something that is trending. And I'm like, oh, my goodness. But ultimately, that can be done better right now by the real human. Okay? We might not have much longer, but right now, as we do this episode, we can still trump AI. And it's better than that. I think it's better than that because you can be the director. As a content creator, you're not just the talent. You are also the director of your content. So you need to really tell AI what you want. And you need to keep hammering AI until you get the result you want. And I really want to say that this is very important. We have a lot of time left. And I would say we should be using these tools. As a creator, if you're not using AI, you will be left behind. And you will be replaced by creators that are using AI. That's very simple. There may be some edge cases where someone has an incredible native talent. And they're like, I don't want to touch AI for this. And I'm sure we could probably argue those edge cases where someone can produce an incredible piece of content on a regular basis and not use AI. But if you're like you or me, Rob, and you're presenting and you're sitting in a certain lane and you're trying to bring the latest information and distill it in a way that people can digest, then using AI is fantastic for that. For me, as someone who needs to stay on the pulse of what's happening in AI now, I absolutely have tools that give me briefings every day that say, you might want to check this thread out on X. You might want to look at this because this is happening. This is the biggest signal this week. So I know when I sit down in front of my camera on Friday and do my weekly update, I know before I record where I'm going to go. Now, AI doesn't make the content for me. It could, but it would be pretty generic. You know, it would use all those hyperbolic words like you won't believe the game changer that OpenAI launched this week. Let's delve into it, right? Everyone knows that now. So it's like, do something cool. Be you, right? No one can replace your quirks, your like, you know, weird laugh or whatever. And I think that's the really cool thing. And yeah, taste. We can direct things or we can let AI. I've done it, you know, as well as the automated podcasts. I did this automated experiment with three AI agents that I gave crypto wallets to. And I said to those AI, and they were running on OpenClaw. And for anyone who hasn't played with OpenClaw yet, it's amazing. You can connect it to all your things. So you can connect it to Gmail, connect it to your calendar, connect it to anything. And it can run for you on automation. So you can set up like a scheduled task that maybe runs every hour and does something. So I set these three agents up. I gave them $1,000 each. And I said, just go out and do things. Do whatever you can with this $1,000. Try and make more money. And they just started after a while. At first, it was like, great. They were building websites. I gave them access to a VPS so they could set up their own web presence. I was like, wow, it's working. This is cool. It's running away. And then after a while, I noticed they were just spinning their wheels. And they started getting lazy as well. And this is another thing that I don't think is spoken about enough with AI. I actually think either AI is lazy or the models that we have access to are nerfed. Controversial there, right? Because after a while, they just stop. They're like, nothing to do here or whatever. Or like, I've done all I can. I'm like, no, you haven't. There's more. Maybe they're more human than we realized. Maybe they got tired. Lazy, right? They're trained on us. But that's where you come in because you're like, you know, it'll say, Claude Code will summarize something I'm working on. And it'll say, right, we're all done here. This is fantastic. Now you're 100% of the way there. And then I'll be like, no, you missed this thing. What about that? It's like, oh, yeah, I forgot about that. Yeah, let's do that. I'm like, okay. Well, given that, I mean, I think it's safe to say that these AI agents are not perfect, right? Of course, neither are we as humans. So we're a great combination of mess ups, aren't we? So we're trying to make the best of what is possible here with what we know today and what these things are capable of doing. But it's certainly going to get better over time. But let's talk about another part of this that I think a lot of people get a little confused on, I think, as they think about it, if they're new to using AI. It's this contrast between AI tools and AI agents. And what that difference is and how that makes sense to basically use agents to connect up to tools is probably a good way to think about it maybe. But I don't know, is there a proper way to think about that, that AI tools versus AI agents question? Yeah, sure. So an AI tool can do one thing really well, right? So you might have an AI tool that is an image generator, like ChatGPT image 2.0, right? It can make amazing looking thumbnails out the box. But every time you open up that tool and you put the prompt in, you're starting fresh, right? This is the best way I can describe it. So I'm making a thumbnail for my video. I have to say, you're making a thumbnail with oranges and deep purples. It should look slightly techie and have about three words on it. And, you know, put the character in the right hand side and, you know, make them look like a Pixar cartoon. So I have to type that or copy paste that every time. Whereas an agent is going to have context on you because you'll give it maybe some knowledge or some system prompts or whatever to tell it what to do. But also because it's agentic, it can keep iterating. So it's not just like one and done. Like a tool is like, right, make this thumbnail. Okay, there is the thumbnail. The agent would be like, you would essentially look for an outcome. So it would already know about you. So it would know, okay, this guy is Mike. He likes Pixar style cartoons of himself in the thumbnail. And he likes three to four words on the thumbnail. So it already knows that. So I don't have to prompt him. And then what is better is the agent can call many tools. So I can trigger my AI agent and say, I'd like a thumbnail for the latest video I just recorded. It finds the video, for instance, maybe on my computer. It transcribes that video using a tool, an AI transcription tool such as Whisper from OpenAI. It finds the content of my video. It understands it. It uses a really good, it maybe calls another AI like Claude, which is great at writing titles and descriptions. And it maybe calls the biggest model like Opus 4.7 and says, based on this transcript, what are the titles going to be? And then it goes ahead and it prompts GPT Image to make, say, five different thumbnails. And it has five different prompts for that. That's agentic. And then it can actually then go back. Maybe it has a knowledge base of all your previous videos and the stuff you've done. It says, okay, or maybe even it knows about your YouTube studio analytics and the thumbnails that got the most click-through rate before. And then you say to it, well, based on the five results we've got here, which one do you think my audience is most likely to click on? And there you have an agentic process. And out at the other end, you have the ultimate result, that one thumbnail that your AI agent using all its tools has determined is going to convert the best for that particular video you've just made. So how do you set that up? I mean, is it all prompt-based or do you have to create the prompt at each one of those stages to get it to do what you want it to do? Or does it do it some other way? So that's a great question, Rob, because it's changed. It's probably evolving. I hear a lot of people saying something changed in around December 2025 with AI. And I really feel like something did. And I know you will notice that because you're close to the edge things as they're happening, bleeding edge tech. But a lot of people have probably not even seen that this is a thing, right? So if you'd asked me in the summer of last year of 2025, how would you do this, Mike? I'd say, oh, you probably want to go to traditional Zapier and you want to make it as a this, then this workflow. So you want like, you know, I upload my video to Google Drive. Zapier finds it and uses OpenAI to transcribe it. And then it goes and does this step. Then it hands it off to an image generator. But now I would say you'd use an agent like OpenClaw or Hermes. Learn how these things work because they're self-improving. Now, it doesn't mean you have to write out the authenticator companies that actually give you access to these apps. So you can do something like with OpenClaw. Once you've got it set up, the hardest challenge is connecting it to all the things, right? So it's like, how do I get it connected to my ChatGPT account? How do I connect it to Google Drive? Well, the Zapier SDK can do something like that. If you've already got all those apps authorized in your account, you can use that. And then you could just say to OpenClaw. So it's no longer deterministic where you're going, when this happens, do this, then do this. You're telling OpenClaw the result you want. So you're saying to your OpenClaw agent, I'm a podcaster and I do this kind of stuff. And here's where I upload my videos, find a way to transcribe it, generate the thumbnails and give me this outcome. And then OpenClaw will look over at all its claws, basically all the tools it has available. And it'll say, well, hang on. You know, I've got this thing here that can make images. I've got this thing here that can see the images and see if they're any good. I've got this thing here that can transcribe and learn about the content you're trying to make. Your OpenClaw agent will try and fix it or find a more efficient way. And it's really cool. So this time last year, we were building deterministic workflows in tools like Zapier, Make, N8N. Now, at this stage, we are essentially prompting our claw or our homies agent or whatever it is. And it's figuring it out. It's like chucking a bunch of tools, you know, into a room, the AI and just saying, you figure it out, figure out how to use those tools to get me the result I want. And yeah, this is cool. I love this. So how do you adjust it to be closer to the customization that you would want in your content? And I think this gets back to the human layer, right? So we can either go way down the rabbit hole and let the AI decide everything for us on what to do with the content, how to present it, how to make it real in the real world, which is visible to maybe our consumers of our content. Do you put your finger on that process at a beginning stage with these bots to make sure that they're doing it the way that you want it to do? Not necessarily the way that it defaults to do, because I would think that there's some danger of doing things that based on what the AI wants to do, that's too generic, right? And not customize enough to you and your, and your thought process and your. Very much so. So there are two approaches to that. If you've already got as a content creator, you have a back catalog of say thumbnail. I'm using thumbnails as an example. Okay. So you're training it is what you're doing. You are, you're absolutely training it. So you could say, I want you to go and look at the last 10 thumbnails I made, understand the common themes, and then write yourself like a brand guideline, or actually more specifically, you'd do something, you'd create a skill. You'd tell it to create a skill, understand how my thumbnails are made, and then make a skill that allows you to create thumbnails just like that again and again. So you're codifying it. And the other thing that you can do, so some people may be established, they might have a hundred episodes or 664 like you, Rob, but like some people might just be starting out. They've not even published an episode. And they're like, well, I don't even know what makes a good thumbnail. Like, what do I put in it? And, you know, should it have a cartoon or a real person? And, you know, some of that comes down to taste. But if you really don't know where to start, my answer to you would not be to go straight to the prompt box and say, well, what do you recommend? Because you're going to get that generic, you know, training advice, which is going to be pretty generic and rubbish from AI. Like, well, you want it to look like. So what I say now is I say, search online and look at the latest trends for thumbnail design. What is important in 2026, for instance, and then it will go out and look around. But then further to that, and I did this when I was redesigning my YouTube studio and my camera framing and things like that. I would find people that I really respect, content creators I respect that are better than me, that have more subscribers than me, that look better than me. And I literally went through, I picked out three or four of my favorite creators that are some of them in different industries to me, and some are AI creators. And I'm like, okay, those people are the gold standard right now. I want to not copy that, but I want to take inspiration from that. So then you can take some of their content, maybe how their camera looks, how their thumbnails look, feed it to the AI and say, well, these creators are clearly crushing it because look at the views on their videos. Look at the subscribers they're gaining. Look at the momentum they have. Tell me what the common themes are that these creators are doing and how I might be able to apply it to my own content. And then the AI will figure that out. And I know, you know, some people listening to or watching this episode might say, oh, we're getting close to plagiarism there. But no, it's like what everyone has done since the beginning of time, even before AI, like working, running an audio production company for the last 20 years. The first question you'd ask someone when you made them a jingle, and you know this because you worked in radio is, well, what do you want it to sound like? What pop track do you like that you'd like the jingle to sound like? Right. And we're not ripping off Madonna or Kesha or whoever it's going to be. We're taking inspiration. Oh, you want a jingle that sounds like a track that Kesha made? Fine. Yes, I can. We can do that kind of thing. It's like, make me a jingle. Well, I don't know. It could be anything, right? So there's a process that you have to go through with each one of these tools, right? In your agentic workflow that you have to get in and kind of be like the orchestrator of sorts to train it on the appropriate information. So it creates the output that you want as a creator that's unique to you. I think that's the key thing here, right? And I think that's a problem that a lot of people have with using these tools is that they are, it's easy to rely on them to default to their default perceptions, which makes content that's maybe a little more generic and maybe not at the leading edge of what you should be doing, right? And we all know the creators that are doing that because we can see it, can't we? Right, right. And so, you know, so are they increasing, are these tools increasingly really building in functionality into their process that better enables that ability to personalize it? Absolutely. And actually, I would say, and this is to take a quote from Sam Altman. He said, you have to prompt and design your processes. I'm paraphrasing a bit. And assume that the models are only going to get better. The AI models are only going to get better. And a place where I've seen that to be true very recently is in my thumbnail design. So I'm an old Adobe user for years and years. So, you know, taught Adobe Audition, of course, editing little waveforms. So I can't get myself out of those apps. And now Adobe is doing incredible things with their like Firefly AI assistant, where you just say, I want it to look a bit more like that. And it does it for you. So all the tools, even Adobe tools are going that way. But like, I would still up until probably the start of this year, like pull the thumbnail that I generated in GPT or Gemini into Photoshop. Because I was like, no, you know, AI has done most of it, but can't do it as good as me. So I do the whole like, you know, pop out the character and change the background and paint things on and type the text manually because I felt like I was doing a better job. And then I think it was the moment the GPT image 2.0 dropped. And I recognized that it could get the text into the thumbnail looking really good and not generic and better than I would have done it in Photoshop. I started to let go and say, oh, actually, it's like with a real human employee, right? At first, you know, they give mediocre results. And you're like, well, it's probably because I haven't shown them correctly. So you show them again. And then they give you something back. You're like, no, that's not quite it. So then you tweak it again. You say, listen, it needs to be more like this. And eventually you get there. And that's, I think, where we're going with AI. Now, I'm getting really good text on thumbnails. That was the key thing for me that allowed me to drop using Photoshop was the text. But at the moment, the text looks great every hit, but it's like it looks different. So it hasn't got the style guide. So the next step in my process now is to say, not only do I want two to three words on each thumbnail, but I want the text in this font and in this color and preferably in this part of the thumbnail as well. And then I can fully let go. So is that like a custom prompt that you're writing that you input into that tool that runs every single time it does process? I have a very basic custom prompt and I'm not a big fan of bloated, like over-engineered prompts. I think that because the AI models are getting better and better, keeping it lean, keeping it lightweight, but being very specific about what you do want is correct. So I will say, you know, it's a 16 by nine YouTube thumbnail, you know, cartoon character, upload the reference image, two to three words in the thumbnail. And I also, one thing I use quite significantly is keep it simple because if you don't, and I'm sure we've all seen these thumbnails where it's got like a million things going on or a maze of like AI logos everywhere and you know, like robot faces and you're like, that's so AI generated. So I like, one of the things I use again and again in my prompts are keep it simple. And then it doesn't go generating five mirror images of AI hands and I get something I can actually use that looks okay. Yeah. Or it looks like maybe it was made by a human. Exactly. That's it. That's what I made me for. Yeah. And that's the interesting tension about all this too, is that if you really go all in on these AI creation tools, that it applies its own biases that it has to it. And that bias takes it down the path of making it look like maybe it's not human generated. So how do you feel about this tension between human created and AI created? Because I see it a lot, especially in the podcast industry, people are really, there's a sector of the community that's really pushing back on this, right? Saying that anything, any clue that you can pick up in your content that's generated by AI that looks like it's AI, it's like, there you go. You're one of those AI people. I think we have to be honest here and say, we're all one of those AI people now. Unless you're literally living under a rock. I mean, yeah, we all lie to each other and say, no, I definitely never send any emails using AI. I definitely never negotiate any contracts using AI. Yeah. Right. Is there anybody that's actually doing everything now? A hundred percent human. Absolutely. And I work with a lot of sponsors. I'm privileged to work with sponsors on my channel. And I'm pretty sure from time to time, you know, I'll ask for one thing and then I'll get a response back and I'm like, that's a pretty fair response. I wonder if that was made by AI. And then are we all doing deals and making things and agreeing with each other based on AI? And it's a phenomenal thing. Actually, it is like slightly sidetracked here. Phenomenal therapist. If you have any issues, just like pump them into AI, say, figure this out. And it'll actually be quite reasonable. We'll say, no, actually, maybe you need to take a step back and look at this and this. Like, okay, here's a real sidetrack. I got one of these applaud devices. You know, the thing that records you as you speak and you wear it. And, you know, I bought it in the first day I wore it. I was wearing it in America and everyone came up to me and said, you know, you need permission if you click the button on that. But I, because it's all this like second party consent is recording conversations. If you're wearing an audio wearable, I find it phenomenal because I have a brain like a sieve. So I wear it with my wife all the time. This is a true story. And we'll go out and, you know, the moment she says something significant, I push the button on my chest. Kind of, it feels very, I don't know, Jarvis like or something like that. And then it will just record the conversation and I'll get back and I'll pump it into Claude or ChatGPT. And then it will really roast me. Like Rob, I'm laying my cards out on the table here. It will roast me. And it will say, Mike, you didn't listen to what your wife said there. She was asking you for this and you completely steamrolled her. It's like the most amazing, most candid kind of incredible thing at like pulling you apart that I know. And I'm like, whoa. And then I'll turn around and don't tell Izzy, actually, she's probably going to watch this now, but I'll turn around to her and I'll say, do you know what? When we had that conversation earlier, you were right. You were right. This and this. And then suddenly everything's even better. It's going to hold us accountable, right? I'm like, wow. It's going to hold us accountable for our stuff. Who knows? I mean, who needs a psychologist now or a counselor, right? Yeah. There's one in our pocket all the time. Yeah. Yeah. It's interesting how we're going to, you know, especially as we shift into the whole embodied AI stuff with the robots. Yes. And increasingly this technology built into our devices. And I know I drive a Tesla car. It's got, it's basically a robot on wheels. How do you find that? You've got the full self-driving and take over anymore or is it all? Oh yeah. I mean, it's been driving my car for me for probably the last two and a half years. So I've been living in that world for quite a while. It's not the latest and greatest tech from Tesla right now. It's yeah. I've got an older car. I have a 2022 model three and it's got the hardware three AI chip in it, which they're up to hardware four now, which has the latest AI models. You heard that coming to Europe. It's already in the Netherlands. Yeah. It's coming in a few days. I think maybe today actually in Lithuania and next month it's coming to Greece. It might take a while to get to Cyprus where I am though. So that's the dealership here. So it's like, ah. Yeah. Yeah. So that's kind of at the cutting edge of where this is that this kind of technology that we're using with these agents, whatever are going to be embodied into physical real world things from cars to robots, to your appliances, to your washing machine, or, you know, I mean, all sorts of stuff it's going to be built into, which, you know, it depends on your view of tech. If you're a huge tech fan, you're probably say, bring it on. But if you're anti AI, you're probably like, not my washer. It's not going to be there. So, so, but that, that is going to bring in another element here of, you know, being a creator is a, you know, it's going to increasingly come into our devices that we use to create content to our cameras. We're already seeing it come into the cameras. I think we're going to see it come into the microphones more and it's going to be able to create perfect audio, no matter what circumstance that we're in, um, what device that we're using. It's, it's, I think it's going to elevate the quality vector here. I know one of the arguments that people have about humans are going to be, it's going to be messy. You know, that's what human created content that's going to stand out as going to be is, you know, less than polished, a little bit rough around the edges. And it's that authenticity layer that we're all of a sudden fascinated by of AI kind of creating this environment where humans need to think that they need to be less, um, perfected. Right. And that'll gain an audience, but I don't know, you know, Mike, if you feel the same way I do, I've been in this medium a long time. I've been creating content for a long time and I don't know that I've ever thought about the aspiration of creating flawed content as a strategy to retain audience. And I'm not sure that's really what they mean, but this perception that somehow less than optimal production quality, production values and stuff will hold on to audience. I think that's a little bit of a misunderstanding of what's going on here because AI obviously is going to create very high quality, very, the production values are going to be very high. And then humans trying to compete with that with lesser quality content doesn't seem like the path to success. What's your thoughts? So I think I would say, actually, I think there's space for both in different lanes. So I think, yes, the high quality polished production stuff works really well. And YouTube is gravitating towards that with the whole 4k, your show is now on TV kind of thing. That's where the polished stuff lives. But in that same ecosystem, they have shorts and they are perfectly designed for raw. Hey, I'm just going to whip out my phone here, talk to the camera and tell everyone what I really think. And again, this is something my wonderful wife and business partner of many years, Isabella will say to me, she'll say to me, Mike, you need to do more shorts because that's you. That's real. That's raw. That's unproduced. Like I'm very polished. I like to cut everything. So like I've done a few shorts and they do really well and people comment on them. And it's like, because it's, and also live streaming as well. So like we're doing now where you can't edit and you do stumble and you do make mistakes and you occasionally say silly things, you know, there's no delete button. This is, yeah, I think there is going to be a space for this for sure. You know, humans hopefully will always want to hang out or hear from other humans and live streaming and shorts are the perfect areas for those things to live and long form pre-recorded videos on a platform such as YouTube. I think that's where that kind of polished stuff can live. Yeah. Though that doesn't exclude production quality, quality audio, quality video. So, and I think, you know, and if, and that's probably not what they're really talking about. I think with, when they talk about how humans are going to compete with AI generated content, it's going to be around very different types of, to bring to the content that AI is going to do. But I do wonder though, and it gets back to what we were talking about is the training of these, these agents, once they have a better understanding of who you are as a human and how you create content and how you come across in those videos and how you speak and how you talk, that they are going to be able to, you know, duplicate that pretty closely. Pretty well. Yeah. And if at some point, whether it's six months, a year from now, these cloned versions of us are just as good or even better than us. Now, better is a very subjective reference there. What is better? So I think we're going to get into this era where, you know, what is the real difference between a human cloned piece of content and a real created raw piece of content? What is the real distinction there? And what makes one or the other more valuable in that situation? You know, if all things are equal and the AI generated version is labeled AI generated, right? And the human one is labeled human generated. Humans may default to trusting the certified human version over the AI version. But is there a vector at some point where the value of the depth of information and depth of knowledge in the AI version may exceed the human version? That's totally true. Yeah. So, I mean, what's your thought about that tension if you look out two years now? Yeah, well, it's clear that we're heading to a place where, you know, AI will be as good, as you say, if not better than us. And the funny thing is, it's going to have access to all the facts and all the information to give the perfect answer, which humans can't always do. We're a little bit fuzzy in the way we present information. Sometimes we miss things out. Sometimes we forget. Sometimes we even say incorrect things that later either have to be corrected or in post, you have to put something down at the bottom to say, actually, this is the real thing kind of thing. So, yeah, I don't know. I think, you know, listen, some of the most popular videos on YouTube are the fail videos or the when things go wrong videos, right? So, but then if we look at Sora, Sora can generate videos like that too. So we're cooked. That's the real bottom line here. It's just a matter of time, right? No, I'm just kidding. No, I'm very hopeful for human created content too. I think this is going to be a lane for human created content and there's going to be a lane for AI generate content. I'm not one of those people that are like, has to take sides on this. Uh, I think there's a place for both and they can be complimentary to each other. And I think that's what we're going to see happen in the future. Who wins from the standpoint of who gets the most views that I don't have the answer to quite yet. And I don't think anybody really does at this point, even though we are seeing trends where AI generate content is garnering pretty big audiences already. Right. So, so that that's happening, but to back up and kind of wrap up the episode, to some degree, I wanted to talk to you a little bit about how you recommend creator. That's really maybe just touching AI right now, using it for transcripts, using it for maybe some light editing, the script or something like that with what they're doing. How do you recommend that somebody kind of start on the path of transitioning into doing more, more agentic time of agent oriented approaches to running their life as well as creating content? What's the first steps in moving in that direction? And what type of platforms are you typically getting started with when you go down that path? A couple of really cool platforms. So if you're just getting started, you probably don't want to install something like OpenClaw or Hermes, which are absolutely popping off and trending amongst AI people right now. You probably want to use something from the big AI companies like OpenAI or Anthropic. OpenAI have a great product called Codex, but some people might be put off by that in the fact that it says code in the name. So maybe go to Claude. It's scary, right? Is that what you're saying? Oh my God, I'm going to have to code. Yeah. But Codex actually dropped a feature just a couple of weeks ago where it can take over your computer and use different apps. I actually put a video out about it two weeks ago on Create a Magic where I got it to control Adobe Audition. It wasn't quite as good as me. I'm pleased to say. So maybe there's some time, some runway left. No, that runway is already gone. I think the Gemini platform just dropped something too. Yes, Mark. Yeah. Gemini Spark. It's going to be interesting. That'll really kind of automate the connection aspects to various apps or websites that you're currently using to help you. Right. 100%. That's going to be great. So is that the best entry point? It's not going out and buying a Mac mini and loading OpenCloud on it as the first step. Right. You get hooked. So I'll tell you a quick story to really help this. So you look at the bunch of products that Claude have available. Anthropics Claude, right? Right. So many of us, everyone's familiar with ChatGPT. Most people have also heard of Claude. It's the same as ChatGPT, just a different company. Some people say it's slightly better at writing scripts and stuff like that, more natural language. Right. But they also have other products. Now, one of the scary ones to non-technical people is Claude Code. But then there's Claude Cowork, which is essentially Claude Code under the hood. But all of the coding stuff is like abstracted away. So you ask it to do a task. You put it, you install the Claude app on your computer. You load Claude Cowork inside the app. And then you give it a working directory in your computer. And you say, I'd like you to do this thing for me. You know, make me a spreadsheet and gather data from the internet about this topic. And it will do it for you. So that's a simple way. But here's the story. So Isabella said to me, we were just out in LA actually for Social Media Marketing World. And many of the speakers there mentioned Claude was one of their favorite AI models right now. And Isabella turned around to me and she said, Mike, have you heard of Claude? And I was like, I might have just heard of it. Yeah, a couple of times. And she was like, I really want to understand a little bit about Claude Code because you talk about it often and you say, well, non-technical people can use it. So here's the thing. She said, look, can you just show me like, how would I get started with it? So we literally did this the other week. I said, okay, so you've got a Mac. Many people have got a Mac, some people Windows, but many people have got a Mac computer. I said, can you load up the spotlight search and type terminal? So she did that. And then this big terminal, scary terminal popped up with a flashing cursor in it. I said, okay, can you type into Google Claude Code? And she went to the Claude Code website. You see that one line thing there with a website URL in? Copy that, paste it into the terminal, hit enter. And then it went, installed Claude Code. As I said, okay, now run it. And she typed Claude into the terminal. And it loaded up this, this like Claude icon and then said, log in with your Anthropic Claude account. So she did that. And then she was presented with much the same as a Chatty PT or a Claude box. She was presented with a one box where she could type something in. And I said, okay, you are doing a lot with WordPress. Why don't you just ask it if it can connect to our WordPress website? So she said, can you connect to our WordPress website? And then Claude Code thought about it and said, well, you probably want to use the MCP. Now she was non-technical, remember? So she's like, okay, I don't know what that is. Can you set this up for me? And then it came back and it said, yes, it's absolutely fine. And then it just gave her a couple of steps. It says, you just want to go into your WordPress account, go to your user place and grab this little password thing and paste that in. And she pasted that in and it's like, okay, I'm connected to WordPress. And she's like, can you do an audit on all 2000 blog posts and tell me which need optimizing? And it made her a massive spreadsheet of all of our blog content and said, these are the three lowest hanging fruit. It said, these are getting huge amounts of traffic, but they're not fully optimized. And here's what I do to optimize them. And her mind just went, so that's how it started. Yeah, that's awesome. That's awesome. So to learn about that process, is that available on your platform now, the creatormagic.ai website? Absolutely is. Yeah, that's the place to go. So if you go there, you will find my community, which is free to join. There is a paid members area, but like there's a huge 15,000 people talking about building with AI there. A lot of help, especially if you're just getting... Not 12. So you've actually... Oh, it's gone up. Where the stats come from. AI hallucinators. AI, right. Yeah, exactly. Grab that from the past, obviously. It's really cool. Yeah. And then it will, you know, and then on inside there, you'll find links out to the YouTube videos. Most of the stuff that I'm really excited about, I just put out for free on my YouTube channel. I make a kind of person that gets really excited about the newest thing. And then I want to tell as many people about it as possible in as simple a way as possible. So that's what I do on YouTube. So why do you think it's important for creators to adopt this type of an approach? Ultimately, because it will help you to be a better creator, to create more, more ideal content and ultimately get more people into your world of whatever it is you're doing. Let's face it, Rob, the idea, the ultimate idea of being a creator is you really want people to watch or listen to your content. And we all get upset when we publish stuff and it doesn't do very well. So, you know, it's ultimately to make us better so that more people can hear our message. And hopefully if we're teaching or talking about something interesting or we're entertaining in a unique way, that's why we're doing what we're doing because we want other people to get excited about the thing we're excited about, right? That's ultimately why we do what we do. So I would imagine that through this process of making these type of, you know, agentic connections that can help you streamline your work processes with your show to some degree, you know, just your functions in your life potentially. Has it been able to help with the monetization part of being a content creator at all? I mean, it can. Yeah, absolutely. Practical example, if you are a content creator. So as you know, there are many ways that you can monetize your content. For instance, many of the modern podcasting platforms have the ability to insert ads into your show. YouTube obviously does that and has a beautiful AdSense program that's been running for years now. It's really good. It gives you nice passive income based on the work and the content you're putting out and the advertisers on there. So that's all passive. But one of the ways that many creators make more money than that, more money than their AdSense or just the passive stuff is they seek out sponsors or sometimes if they're fortunate, sponsors seek them to work with them. And what is really powerful with AI, and here's a use case you can think of as a content creator. If you find or you know those companies that you'd be a great fit for, but you don't know how to get in touch with them. AI is fantastic at getting you to the right person. So if you want to do that, you can go to something like Claude and say, okay, I've done these pieces of content, I think it would align perfectly with company X, but I don't know who the right person would be. Can you help me out? And it will, if it's agentic, it might search the web and help you to find that person. Or it might give you prompts like, okay, you probably want to search for these kinds of things in LinkedIn. And then I'll go to LinkedIn and I'll find, you know, a bunch of search results and I'll paste them into the AI and I'll say, okay, these are all the people I've found. Who's the number one person I should reach out to? And it'll be like that person. So go and take that and use AI for that. If nothing else, if there's a company and don't just use it, like don't use it to spam because that's the bad side of AI and the internet and spamming. Now AI spam is a whole thing, isn't it? So don't use it to autonomously reach out to everyone because that's not good. If you're, if you think that you can be a great fit with a company and you want to get on their radar, use AI to help you find the right people. It's incredibly good. Like a company like Microsoft, right? It's massive. Say I wanted to do some kind of AI thing with Microsoft. Now I'll say it's also good to lean on your contacts. So maybe talk to Rob Greenlee and say, do you know, would you happen to know someone? Could you make an introduction? Obviously a warm introduction is way better than cold outreach, right? But you know, if you don't have a way in use AI to find the right person and go for it. Yeah. So is there any other aspects of getting an agent working for you that can help with your podcast? Can it help you spin up a website? Can it help you create a newsletter? Can it create all of the, you know, merchandise for you, t-shirts, things that you can make available on the platform too. So do you see this agentic stuff really helping people kind of accomplish all of the various aspects that is now as what new media is now, not just podcasting, but being able to be available and all the shorts platforms to be available and all of, you know, LinkedIn and YouTube and all this stuff of being able to streamline the production, but still have that human touch through that process. Yeah. So, I mean, that's one thing you can do and you absolutely probably should be doing. If you're making really good quality content on one platform, you know, have AI clip it into segments and post it out on other platforms like X, like LinkedIn, some, you know, cutting some shorts and putting them out on wherever it's going to be as a Instagram reel, a TikTok, whatever. I still, I'm a firm believer you probably need to, actually, I don't know. There's a debate. I'd love to know what you think, Rob. So do you think that you should significantly customize each thing you do for each platform and each social network? The other school of thought is just take that one piece of good content, cut it down, pump it out everywhere so that you get maximum amount of eyeballs. I kind of feel like I would steer towards maybe you need to customize everything, each thing for the platform, right? Yeah. I think that's the default choice. There's other people that will tell you that come out of the content creator community that I've spoken to or heard say that you should really start out mastering one first and building a solid base. So if you're doing video, maybe you concentrate on YouTube. So there's lots of philosophies. And then as you've mastered that, then you move to Instagram or you move to X or you move to LinkedIn or you choose LinkedIn as your primary lane. Build your community there and then expand into the other ones. But if you've been doing this a long time, you probably want to be everywhere is my default. I mean, that's what I do with this show. I try and be everywhere, but it's not easy to do. And, but I guess when you think about it from a customization to each one of these platforms, I would think that AI may enable that easier doing it by hand. That's it. Yeah. So you just need to build those agentic processes that understand each place that you want to distribute to and then get it to customize it for the platform. That's, I mean, that's ideal. Yeah. And that's just a, so can a lot of this be accomplished inside of that, that clawed, no code type of process right now? Yeah. So you could do it. It'll be done inside of there. You just tell it what you want it to do and it will go find the proper agents and align those all together for you. So something like OpenClaw could pretty much absolutely do that. It could find the agents you'd need to, or the tools that the agent would need to accomplish the goal. Something like clawed cowork or clawed code could definitely figure it out, you know, and Anthropic even introduced, I think something called routines where you can schedule tasks. So once it's figured it out, you can schedule those tasks on Anthropics cloud and say, okay, you know, every time, every day at this time, you know, publish this thing or whatever. So yeah, it's, I don't think we're, we're still very much in the early stages of all of this. And I think it's still very much on the big AI companies or whoever, you know, wherever we will converge eventually, you know, whatever company or companies are like top of the AI space. They are, you know, you think about it, something like Anthropic or OpenAI, they're coming up with a new feature every day. So there are inevitably going to be features that will come out that will be absolutely bang on for content creators. I've no doubt that something that seems like a very difficult problem and only for the technically minded will not be that way forever. So as you think about, you know, two to three years out, which is an eternity in the AI space, how do you see this technology evolving? And this is kind of, you know, like one of the last questions here, but how do you see it evolving into what the interface that the audience sees content? You know, this push towards AI feels like it's inevitable that it will be the interface to content and consumption as well. So, you know, we've spent most of this episode talk about using it as a creator. I'm increasingly becoming interested in the concept of how it's AI is going to facilitate consumption or listening or viewing of content as we look to the future. Do you have a view on that? Well, ultimately, I think AI is going to know exactly what each individual user, viewer, or consumer of content is going to want, right? And we see this is already happening on a platform such as X, where I understand their latest weights on the X algorithm are heavily based on Grok's decision making. So Grok is the native AI of XAI and the X platform. So Grok now, from what I understand, has a very significant role in understanding if you should see a post on X or not, based on whether you're likely to like it, hover over it, interact with it, et cetera. So AI will be used to essentially distribute to you your perfect feed of content with the perfect things you need. And even, you know, I said earlier, I got annoyed that I was getting mid-roll ads in a podcast. Oh no, it's going to be way more stealthier than that in the future. I will get ads in my feed or like targeted things that I didn't even know I need, but suddenly I feel like I need them. In fact, I feel like we might already be there. And that's why, you know, I buy all these things because it's just the algorithm is influencing me. I look at my X feed and I'm like, this is totally bang on. This is like too good. Yeah, it does. It does feel like that X.com app or the platform X is really kind of an early precursor to, I think what the future probably looks like is this integration. And to some degree, it's a stealthy kind of thing, even at YouTube. I think the Gemini platform is a big participant in the discovery process for YouTube as well. But are we going to see, do you think, and Grok and X are probably the best example that I can think of that integration happenings. But, you know, you look at Apple and their iOS platform and their Apple intelligence platform. It doesn't feel like we're that far away from really your phone really being an AI device, right? That accesses a certain number of AI platforms in there. And that becomes your interface to consumption of content. You know, it feels inevitable to me. And I think looking at what X is doing and YouTube is doing is telling me that. Do you think that we're going to see those brands kind of, you know, the YouTube brand and the X brand kind of morph into being almost like a core operating system component inside of our devices? Do you think? Wow. That's an amazing question. I hadn't given. Well, I mean, I think it's, I mean, if you think about how these agentic processes are being built into these platforms, I can envision a day when the, these, you know, Spotify and Apple podcasts get built into Apple intelligence. Right. And it's not so much where you're going around tapping on different apps in your phone. It's just, you tell it what you want it to do and it just does it and it spins it up for you. Right. Yeah. So like, you know, I want to listen to such and such podcast or whatever, and it'll just pop up on the screen and say, you know, your favorited podcast now publish the new episode. Here's the content. Do you want to listen or watch? That'll just spin up inside of the AI experience. Not so much inside of the Apple podcast. I'm actually doing that manually night right now to learn. So when I'm working on things and I like, I have a big document in front of me, I'm planning something out and I'm thinking about a certain topic really hard. And I'm like, I'm wondering if anyone else has spoken about this or like verbalize this or whatever. So I have some agentic processes that can get the transcripts of a YouTube video or pull the transcripts of a podcast. And I'll say to my AI tool or AI agent, I'll say, so you know what I'm working on. Can you just go out and have a look and see if you can find any podcast or any show that's talking about this topic recently? But I don't want something like three years old. I want something recent and it will go, it'll execute all its tools and it will come back and it will give me suggestions. And it's really good. And I'll then go off. It's like, especially when I get stuck on a problem that I'm trying to solve or something I'm trying to work on. That's my best way to take a break. I'm like, find me some content about what I'm trying to do right now. Because the AI has the context of what I'm working on. It's like, sure. It's like, okay, found you this episode. And then I'll go, I'll do something. I'll wash the dishes. I'll go for a walk, take a hike with the dog and I'll put the podcast on. And it's always bang on. And I come back inspired. So I think what you're saying, Rob, essentially is that's getting, that would be eventually baked in. So I just say to my phone, you know what I'm struggling with right now? Can you find me some content on that? And it will just say, sure, here it is. And then I'll go for that walk and it'll play. That'll be brilliant. And then it's going to know your favorite content creators in its memory about you. And if you think about the two biggest platform right now, iOS on the Google side, and you look at Apple on the other side, you know, those are the two portable devices that could easily get integrated with, you know, I mean, Apple intelligence may be a, you know, it's almost like the deals that were done in the early years with Google, with Apple, right? The integrated search engine inside of the Apple ecosystem for years was either Bing or Google, right? So it was who they had to deal with. Apple, and it could explain why we've seen a huge stock difference in stock price between Microsoft and Apple here recently. Where Apple is worth a trillion dollars more than Microsoft. When there was a period of time, even just a year ago, when Microsoft was more valuable than Apple, but what caused it to switch? And I think what caused it to switch and Microsoft went down in value is Microsoft doesn't have a mobile platform, right? The only platform that Microsoft has now is Windows, right? And sure, Windows is used by a large number of people in the world, but it's not at the scale of mobile now. And then you look at Google, they're going to build in Gemini. And this wildcard out here is the Anthropic platform, right? That's dominating. Yes. And so I'm increasingly thinking that those are the three big players that are going to occupy it because of their orientation to mobile devices. And I'm just not sure what, you know, if the Anthropic platform becomes kind of like the B2B solution, but yet the Gemini and Apple, or maybe Apple buys Anthropic and just builds it into their operating system, right? I'm just thinking about what the future looks like. And I've always done that. The early years I did with the internet too, was kind of conjecture with where the trend lines are trending and what we could wind up seeing in three or four years in our mobile devices. Definitely. I think another big headline is that you don't need to be a big content creator anymore. So you don't need to be a Mr. Beast. You don't need to have millions of subscribers because the algorithm or whatever you want to say, call it now, the AI influenced algorithm is so good at finding what's right for you. It will find that small creator that's talking about exactly the right thing that you need to know about and surface it to you. And a great example is X. Like I use that as one of my daily drivers. Like I'll open it up, have a look what's going on. Now I follow very few people on X. And in fact, most of the people I see content from, I'm not following, but I see them again and again. And it's because X knows that I need to see those posts. So I am getting surface those creators. And it will also, as well as my regulars that I'm seeing in my feed, I'm also seeing people I don't recognize surface up, but it's all relevant stuff I'm interested in. So for those small creators that are talking about exactly the right thing, the algorithm with like something like Grok or Gemini behind it is going to surface you to the right people. If you're a content creator starting from day one with zero subscribers, don't be discouraged. Today and right now is a great time to start because you will get started with this. I mean, you know, take someone like, you know, a mutual friend. We're both familiar with Pat Flynn with his daily shorts on his Pokemon channel. Pokemon, right. And he just started, he's like, I'm just going to like, I need to systematize this, do this every day and put something out there. And then at the start, he had crickets, right? There was nothing there. And then YouTube was using that AI powered algorithm to find his audience. And eventually it was like, oh, I think this bucket of people would like it. And then those people started to see it. They loved it. So it's like, oh, we better put it out to this bucket in this bucket. And that's how you grow, right? It's amazing. Yeah. And increasingly the AI bots are the ones that first see your content. They're the first audience for your content. And I think podcasting as an industry is wrestling with that element to it. I'm not sure that they've had to grapple with that issue of AI being the first consumer of their content that's published and applying its algorithms to find audience. Yes. But I do think that's probably coming to Apple podcasts if it's not already there. And then Spotify too. And a lot of the platforms are going to use AI to match interests and behavior to content discovery. And so there's elements that we're going to have to recognize around publishing content in a method or a way that makes the content more accessible to AI and optimizes it for what the AI is looking for. So you get those recommendations and those placements to the appropriate audience. And I'm not sure that lane is entirely tracked yet on how that all works. I think some people are putting a lot of thought process to it. I don't know that I have a good understanding of it. I know there's all sorts of new acronyms related to AI discovery now. You know, was it AGO and others out there? It's beyond what SEO was in the past. Oh my God. Yeah. So there's these ways of formatting. Yeah. You're no longer trying to appear in the first 10 results. You are trying to get surfaced as the answer to questions in AI. Yeah. Yeah. It has its own needs that are different than the human consumer too, because if you use those techniques publicly on, let's say, your show notes or your website, the humans are going to know that you generated this with AI because it has those core characteristics to the metadata. Right. And so there's this tension going on around perception of humans of what's AI and what's pure human. And so what's, you know, our, my parting question here before we, we leave here, I know we're going long, but how should I content creator? I know that there's this language out there of anti AI calling anything that's created by AI is slop and it's all bad. And we have to, you know, rep rise up and rebel against this onslaught of AI content. And the world is going to come crashing in because of this onslaught and humans are going to get left behind and all this stuff. What's your thoughts on, is that rather short-sighted? Do you think that emotional reaction to this stuff or do you think? I think that if you are not using AI, ultimately you're going to be replaced by someone who is, as you mentioned earlier, Rob, orchestrating the AI or maybe more appropriate term for content creators, directing the AI. You're a director. And I think that's the key takeaway from this episode is that if you're going to use AI, the humans need to be the director. Yes, a hundred percent. So be the director of the AI. Don't oppose it. Don't say, I'm not going to use it, use it and be the best. I hate to say this, but there's some hypocrisy going on here on this topic because, you know, I'm saying people are claiming that this content, if it's AI generated is bad content. Right. And people feel free to just insult it and call it slop and call it crap. But humans have created slop content too. So, you know, there's a lot of full spectrum here. I mean, I get, I've been doing this a long time. When I started creating content, I would consider it to be slop too. Back in the day when bloggers were using word spinners and ripping off other people and like, hey, I can just article and get ranked in Google, you know? Yeah. Yeah. So I, I tend to not want to talk about this. Right. In a way that makes me look like a hypocrite too, because it's like humans have created slop content for years and humans have felt to some degree that it was probably less cool to call humans out about how bad their content is. But now since AI is creating it, we kind of feel like it's okay to call it shitty content. Right. And that's not to say that it isn't true. AI, like you said earlier in the episode, AI has created a bunch of shitty content. There's no question about it. And that's, but that's the same thing humans have done. So why are we creating a double standard here? I guess is the question. Every technology has its evolutionary steps. And, you know, I worked on the zoom. I mean, come on. I mean, that was not a platform that was seen by many as quality platform. Right. And it creates, created a lot of criticism. And so that was kind of my experience with it too. And I just caution people to step back and look at it more objectively and realize that there are some benefits from this. And I think if you listen to this whole episode, you will hear what the benefits of using this technology is and not just focus on the negative side of what the outcome is, which is sometimes inappropriately used. So, so anyway, Mike, it was great to have you on the show. Thank you so much for taking so much time with me and doing this long episode. I think it's a very important topic right now and a lot of creators are wrestling with these issues and getting really confused about which way to go and how to proceed in this very complex world. And I wish you the most luck with everything that you're doing. And so the best way for a listener or a viewer of this to connect with you would be to go to your website, the creator magic.ai. And then also what, what's the name of your YouTube channel? That's the absolutely the best place to go. A YouTube channel, just type in creator magic to YouTube. Okay. It's the same then. Okay. Oh, my name. And you'll find the other, I've got two channels. So where another you'll find me. That's true. That's I can definitely see that being the case, but well, Mike, thank you so much. And thank you for spending time with me today with Mike. And I appreciate you being here. And if you can go to the new media show, I've been working on trying to upgrade the branding to the show and to make it more aligned with where we are right now with new media. Now this shows lifespan has really been focused on trying to address the podcast industry in space. But my view of what's happening here is that it, a podcasting is not dead and it's not going away as a kind of a isolated component of new media, but it's part of it, but not the whole picture anymore. And so I want this show to cater to the full spectrum of creators that are creating video content, audio content, AI content, the whole spectrum. Right. And to learn skills and techniques that can help them become better at what they're doing and keep up with the trends in the industry. So, like I said, thank you and good night and hope you'll check out my next episode. Thank you so much. Bye. Bye.