AI agents are taking over: what is vibe marketing?
The Factory Floor · Episode 8 · August 15, 2025 · 50 min
Transcript64 sections · auto-generated, so expect the odd wrong word
All right, ladies and gentlemen, welcome to the factory floor. Uh, I am joined by my companions in business and friendship. We have Corey Eugene Hayes Haynes. That your last name wrong, dude. I've known you for like 12 years. >> Yeah, Eugene. >> Cory Lis Hannes earlier. Yeah, it's got to be you in there. >> That was kind of cool. >> You listen. >> That's a common theme throughout some of Conversion Facto's design assets is I will always use Cory's name, but it's extended by like >> 12. Yeah. It's like Corey Allen Bartholomew Mule. >> Oh, yeah. Yeah. Yeah. >> Edgar Haynes Esquire III. >> We're very professional at Conversion Factory, by the way. Um, so we got Corey, all those names, Hannes, and then
we also have Zach Stevens below me on my Zoom meeting call thing that is joining us. Um, the thing is today we're talking about something that Zach and I apparently don't know anything about, which is perfect because we're going to let good old Uncle Corey walk us through. We um, we're going to talk about Vibe Marketing today. Now, I've heard of Vibe coding and so I'm assuming it's something to do with a vibe coding. Yeah. But not vibe marketing. >> So, >> can you just can you take us on a journey, Cory? Is that possible? >> Nick heard about vibe vibe coding and then he made it so that that word doesn't show up in his Twitter feed because he immediately excluded it.
>> I need you to tell me if I need to do the same thing to Vibe Marketing if it has AI involved. Yeah, probably. Probably. >> All right. Thank you, sir. Corey. >> Well, Zach, have you heard of Vibe Marketing? >> I hadn't actually until you brought it out. I'm assuming it follows, you know, like Vibe Coding, Vibe Marketing. I'm I'm assuming there's a some kind of similarity that's in there where you're just you're going for the vibes, not like the actual technical feasibility of whatever it is you're doing. >> Uh-huh. Uh-huh. Uh-huh. Okay. I think this will be a fun discussion. Um, and it's not actually not that surprising that you haven't heard of it because, you know, everyone's Twitter feed is
different. So, like my I feel like my Twitter algorithm is very in tune with like breaking news and like whatever like the new Zeitgeist is of what's happening in tech. So, I feel like I was like one of the first to hear about Vibe Marketing and sort of already knew that that was probably going to happen. Um, >> and it's always funny when these terms come up because you never really know like where it came from or who started using it first or who popularized it or why we all accepted that it became a thing. But yeah, to to back up for a second, vibe coding is basically what people say when they're not really a programmer, but they're using AI so much that they're able to build apps and they're able to program and develop and
write code purely based off of prompting. So like you're not really there's no strategy or art to it. It's just so like in cursor or in wind surf or whatever uh you know replet vzero bolt lovable like there's all these big um app builder platforms now or idees that have agent capabilities within them where you're literally like um you know build me this app here are the requirements go and then it starts going and whatever doesn't work you're like this doesn't work make it work fix it there's like there's no art to it, >> rhyme or reason. >> It's It's so like brutish, you know? you're just in there like there there's a meme of uh this like gigachad dude and he's I mean he's insanely buff and he's
like working at a standup desk and he's looking at his computer and it's like you know the text is like what he's writing into chat GPT and he's like build me this app make it per perfect no errors do it or you die you know it's like this whole like people are so savage with their prompting so he had vibe coding now let me just say that I was was a vibe coder before vibe coding was a thing honestly. >> Oh yeah. >> That's actually not a good thing. That's pretty embarrassing. But like literally when I learned to code two years ago almost in August of 2023, I was at a coding boot camp at Ryan Culp's house and I had just started coding. I just downloaded my first coding IDE. Just started using a
terminal. Didn't know anything about how anything worked. And then someone said, "Hey, there is a a fork of VS Code, one of the biggest coding IDEs there is that has chatbt like in the sidebar in the IDE, so it can like actually see your code. You don't have to copy and paste between VS Code and Chat GBT." And I was like, "Well, that sounds great." Because that's what I was doing, and I hate copy and pasting things back and forth. So, I downloaded this thing called Cursor. And now like cursor is what everyone uses to vibe code. So I was doing that way way way before it was cool where it was like still chat GPT or uh GPT 3.5 turbo and you know it was writing the worst code in the world but I was just clicking
accept and then feeding it the error message and then having it fix it and then you know you just keep going like that. So that's where they say like you're vibing because you're not really paying attention to the architecture of the app. You're just prompting until it works. >> Um, so does that make sense on what vibe coding is as like a starting point? >> Go ahead, Zach. Well, it just seems like Okay, so the the context is that you're really they say vibe coding quite literally where it's like I'm just trying to get the vibe like and have it get this I I use that term frequently during design projects. It's like h this is like you know it's off brandand the vibe isn't quite right with this particular thing
and it's not necessarily a technical judgment. It's a emotional subjective one which is what it sounds like. It's it's subjectively >> subjectively pointing out whether something does or does not work in the context of coding. >> Yeah. It's sort of like um >> maybe the difference between like drawing a sketch of a house and making an architectural blueprint, right? Like whereas programming before was like you're making an architectural blueprint of these are the exact measurements. This is what's technically sound. this is what is up to code so on and so forth you're just sketching and you're those sketches become reality though um and uh you know so so vibe designing okay so that the reason why they say
like vibe coding is because the AI is so embedded into the workflow and it's so quick to actually do things that you can just kind of take this more spontaneous loose chaotic work style approach. You don't have to be so scientific and precise and technically sound with everything. >> You're just vibing. You're just prompting and seeing where it goes. And it's a very like windy journey to an end product. >> Okay. So now I I really want to hear what does this look like for marketing? >> Well, okay. Hold on. So the just so I understand the putting the word vibe in front of vibe coding is more like I'm chilling, I'm vibing, this is easy. I can just and it doesn't it's less like >> it's like something about the coding
that makes it vibe coding. It's just like no, I'm just chilling while I'm doing it basically. >> Okay. >> Yeah. It's just an easy breezy >> carefree cover way of >> coding. It's crazy how not understanding I am of the term vibe coding and how popular the term is. Okay, >> keep in mind it's popular within the >> you know echo chamber of 100,000 people in tech Twitter. >> By the way, it should be noted that you should we should all use the word zeitgeist as much as possible in as many situations as we possibly can because it's such a good word. >> I love that word. >> Okay, tell us about vibe marketing. Okay. So with Vibe coding, right, we take an established workflow and process and then we put so much AI into it that
we don't have to care as much about how long that process takes or all the technical bits or the exact way that we get to the end product that we want. Right? So now the exact same thing is happening within marketing where we take a process like writing a blog post for example >> and maybe before writing a blog post would involve you know a couple of hours of doing some research on Google reading a book trying to find some primary sources maybe running a poll on Twitter emailing a couple of experts and friends so on and so forth and then writing an outline and then like filling out that outline design into a draft, sending it off to an editor, sharing with the team for feedback, adding some designs, uploading
it to a CMS, publishing, then promoting on Twitter or so on so forth. Now, with AI, every one of those pieces of the process can be expedited and automated by like a hundred times. So instead of going and doing some research on Google, you just have the AI agent go and scrape, you know, you do like the the deepseek or the chatbt deep research and you say, give me a summary of all the primary resources on this topic and then it's going to give you all the facts and then you say, Chad PPT, hey, build me the outline. Cool. Make a couple tweaks to the outline. Now you say, hey, now write it. And then you can have Chad PT go and create some illustrations. You can do some sketches and turn them into
some stock images or into a diagram or into a prototype or a mockup. And then once you publish it, you can have all these automations that then go and create individual unique posts for Twitter, LinkedIn, so on and so forth. You can even turn, you know, for example, what we're doing right now, we need to do some more buy marketing with our process. We could take the transcript of this podcast and then have AI go through a couple of different processes within an agent to turn it into a blog post and then turn that blog post into some social media posts and turn those social media posts into something else so on and so forth. Right? So vibe marketing is basically how do we take AI and use it to expedite
and automate all the things that we do to make it essentially like a spontaneous process instead of a super systematic >> scientific >> drawn out lengthy process. Okay, I have a question then. As a a marketer and someone who is up with the zeitgeist of what's happening and knows, you know, reads a lot of blog posts and a lot of Twitter posts and a lot of these things. >> Do you feel like you can see like this isn't that person, this is AI, like I can I can see that this blog post or this tweet is like distilled from this? like you can kind of you know like it's just like um you know any like article now I feel like sometimes I can read I'm like you didn't write this I could tell do you have that with all this marketing
material do you feel like you have >> buffering buffering I think I got the gist of what you said >> okay go ahead >> um can I tell yes you can absolutely tell here's the scary part though is that it's getting harder and harder for me to tell and and recently I've been seeing people's uh tweets that I for sure thought weren't created by AI and then I see one thing that kind of tips me off that it actually was created by AI >> and I'm like wow like that basically almost got me and like it's going to be impossible to tell really really soon. Um >> and here's the thing. I don't think it's going to matter at all. I don't think it's going to matter. No one's going to be able to tell. No one's going to care.
There's going to be a lot more noise and there's increasingly more noise every single month, especially on social media because creating content has gotten so so easy. There's even noise like now every tweet that I write or that I publish, I probably get two to five automated AI comment replies. >> Mhm. Sometimes I also can't recently I can't tell if those are AI or not but then because they comment on every single thing that I do I realize that they are but actually what they're saying doesn't really read like it was written by AI. Um, but again, this is just kind of the inevitability of AI is with marketing, it started with simple things like, help me rewrite this headline, and now it's like, oh, I can
just feed it a transcript of this podcast, and then in 5 minutes, I'm going to have a published podcast, podcast clips, blog post, social media tweets, so on and so forth, all published and scheduled in the right places automatically. and I'm going to trust that the output is good because it's gotten to that point where it's basically indistinguishable from what a person can do. Um, and so vibe marketers are sort of like the new growth hackers if you think about it. Growth hackers was were cool like 10 years ago and growth hacking was all about building in or taking in the processes and tools uh developers and then using it for a marketing use case. And now the same thing is happening with AI where people are building their own
scrapers, building their own research assistants, building their own uh agents, building their own X, Y, and Z that really really speed up the process and allow you to do things that you couldn't do before or as quickly before nearly as much. Do you feel that the I mean I feel like this is happening with this happens with every democratization of uh particular skill set that the you already mentioned the amount of noise keeps going up but do you think that for like in the hands of a really good marketer this sounds super effective right like okay do you think that in the hands of an inexperienced marketer it's a it ends up becoming a noisy drum set in the garage >> with someone who doesn't know what
they're doing. >> Uhhuh. Um, yes. But here's the thing is that I think that because there's so much noise and there's only so many seconds in a day and only so much real estate in a social media feed that all the bad noise ends up getting drowned out. Actually, >> um, it it we're just replacing manually written tweets that are filling our our timeline with AIdriven tweets that are filling our timeline, but they're basically equal or or better quality. Um because everyone else that's just using it for really bad if you're writing bad tweets, it's still going to get bad engagement, right? If you're writing a blo a bad blog post, it's still not going to rank in Google. If you're publishing a bad podcast, it's
still not going to get listened to, right? So if the quality isn't there, it's sort of like as if it never happened. But for everyone else now, we're making more better things faster. Um and and so you're able to do more with less. That's really the whole theme of AI, right? It's like you're able to do more with less, produce more with less people, less resources, less time. >> Okay. I was going to ask you about that cuz that that was kind of my view is that you could do like there's way more output. The the quantity is like huge. You can do as much as you need or want at any given time, you know, in all these different skill sets or industries. But what hasn't fully caughten up is the quality. So like
quantity is through the roof, but quality is like slowly increasing. Like before you could easily pick out like obviously AI terrible and then like now it's kind of like is that AI? I don't know. >> So the quality is like catching up >> and the quantity is always going to be insane. Like you can do a ton of stuff but the quality is just slowly climbing to the point where it's like dude what if the quality is like way better than we would put together and you can do 10 times more every day. >> That's going to kind of like ruin all of our lives. like I don't want to I don't want to go on Twitter. Like I won't want to look. >> You know, >> there's definitely going to be some things where it's like
>> if you can prove or if you can really go the extra mile to show that like, hey, what I'm doing is not by AI or at least it's not obvious. >> Human validation, >> yeah, then people kind of like give it an extra nod. They're like, okay, this seems to be something that I don't feel like it's going going to try to trick me. But I think there's a lot of things that people don't care about. for example, ads. Um, like just recently, um, as of like a month ago, since uh, Chad GBT released GBT40 image creation. >> Mhm. >> I mean, it's like a mile better, like 10 times better than the last model at image creation. I'm looking at some tweets in front of me. I'm not going to show it cuz everyone listening is not
going to be able to see it anyway. So, I'll just describe it. But um people have been showing a lot of these ad mockups where you can just give Chad Boutique a picture of your product. Let's just say it's a lipstick for example. And um you can be like, "Hey, this is my product. This is what it looks like. Here's a couple angles of it. Now, do some mock-ups of this within models's hands on the beach, uh, with this kind of backdrop at sunset, for example, and it'll create a really, really hyperrealistic, amazing looking ad within a couple of seconds. Now you don't have to go spend tens of thousands of dollars to hire a model, take a boat out to an island, bring all your professional photography equipment, rent
it out for the whole day, buy people's meals, so on and so forth. Even for software, I mean, it's like maybe what would have taken a designer, you know, a few hours for a product UI abstraction or illustration. Like I know it's Zach, you can tell me more what you think about this, but like even though it doesn't feel like it's there yet, it's going to be soon. And there's just so many use cases where it's like, well, maybe what would have taken literally a week or $1,000 or $10,000 to do now is just one prompt away. That's pretty amazing. And again, it allows you to work so much faster. That's why this whole vibe marketing thing is kind of catching on because people are like, "Dang, if I'm a DTOC
brand or I'm a solo founder of a software product or I'm a consultant, like now I'm a one oneman marketing machine. I don't need a whole team to do all these things. I don't need to spend uh $100,000 a year just on influencer marketing. I can just make an AI influencer and then feed it my products and you know so on and so forth. >> Yeah, I think specifically around the concept of images for DVC brands, I think it's crushing it. like it. It's a super powerful tool that's ready to go, assuming that you're able to get at least a couple, you know, uh, plain white background images of your product where it can then assess the like the dimensions and all that cuz like I even took something from um, Athletic Greens
website when GPT40 first got released and I they have this uh, it's a set of all the different things they had like the bag, the canister that it comes in, um, a spoon, And then like even some like sprinkled AG1 dust on the like the white countertop that was on whatever. And I said take this image but put it into a moody lit kitchen countertop and do all that stuff. And I was really impressed by the results for that. There were some things that it still hit the uncanny valley like uh okay the spoon is like sticking out of the closed um canister which isn't going to happen but easily fixable with Photoshop. Um, and the it it does it carries the like raw material a very far away before you have
to come in and say, "Okay, I need to make some adjustments to actually make this real." With software stuff, I feel like it still isn't doing it right. And I've noticed this with anything 2D actually. Like I I feel like just 2D vector graphics, it has a really hard time creating things that are coherent. Like I've even seen the Um, you put this person in our Slack channel. They were doing like brand style guides, like just a a ton of them. And I was looking at them and I think that I I think I put together a video for it, but I didn't get to finish it where all of them were garbage. I was like this. I was like, you picked inter as the primary type face for everything and it didn't even get that right.
There's no, it was very homogenized and it I don't know why, but like it just it has a hard time doing anything digital. Like if there's a digital interface that it has to then abstract and turn into something else, it's it sucks at doing it. And same thing with like creating brand style guides. I was like, "Yeah, but like those icons, they're not even cohesive. I don't even know what they are." Like you just have this circle with a random line in it. It's not a a usable icon. >> Yeah. >> Uh so I guess it's like it's there. Thankfully I think image generation for software stuff is still a has a moat around it as far as creating things that are actually good for the time being. for now for the time being
>> because let me let me tell you why because um every time there's a new model that comes out it's because the team uh developing it you know whether it's openi anthropic so on and so forth they are specifically training that model to be better at something and so for example when you they came out with uh with GPT40 that was a new model just to improve the image generation so for months they've been fine-tuning ing and tweaking for this output. Now, when they put their mind to 2D abstract, you know, concepts and diagrams and um things like style guides or UI illustrations, things like that. They're going to spend a couple months, they're probably even doing it now, and then they're going to come out with a
new model. It's going to be, like, hey, this model is great at this now. And we're all going to be like, how' they do that? It's the same process every single time. They're training on something, some output, some really specific thing that they're trying to get the model to be better at. And it takes a lot of iterations, like literally billions of iterations. And it takes time to get there. And it takes some essentially prompting on their end to try to make the model better at those things. And that's why we're also starting to see a divergence in the models and what they're good at where you see like you know >> versus 40 right >> right so so now we can sort of like toggle on and off these features of a
model where it can take some extra time to go do some research or it can take some extra time to go and get every little pixel right or it can take some extra time or maybe that model is just built for speed you know we have like I think it's like 40 Mini where then there's like 03 something or other. Um they're all going to have these little flavors and tweaks and there's going to be a model specifically that's really good at whatever we think that it sucks at today. >> You know, a year from now, a couple months from now, I don't know when it is, but >> I'm for it. I'm 100% for the empower uh giving people the opportunity to do things that they didn't have the capability to do before. Particularly
people who are already doing the thing manually and I guess that's like the where where this goes hit like if you know how to do all this stuff manually then my gosh like your workflows and your ability to dictate what is and is not acceptable for these things is going to make your prompting so much more efficient. and then you'll be able to generate at mass scale because I do think like if we were like like I said if we were a DTOC marketing agency and we needed to do stuff like like you said a lipstick brand or a you know like a soda or something like that that stuff man I would be over the moon excited with all of the things that we would be able to create and how quickly we'd be able to create them and then how
>> how much more creative thought would be the uh the value that's created rather than the time that gets put into seeing the idea manifest. >> Mhm. Yeah. Um can I ask a question? Go backwards a little bit and ask a question. >> No. >> No. >> Okay. All right. This has been the factory floor. No, I'm just kidding. Um Okay. We started talking earlier about like human validation. So, like the um you know, like at some point tons of stuff is going to be made with AI and then we're gonna have like a special premium feel to things that are made by humans. Um it kind of reminded me of like, you know, coffee in the US used to just be like a like you go in, get your coffee, and you leave or you like go
through a drive-thru at McDonald to get your coffee. And then Starbucks like kind of was like, "No, we're going to have a cafe and you like can come sit down and it's not like this like industrial like get your thing and go." Um, they wanted to make it like an experience. Um, and people had a lot of this type of stuff after CO. It was like after co the like service and experience economy is going to boom again because people miss like just normal human being things and like this this um experience uh versus like the you know get it and go type mentality. Will we have something like this for AI produced products and services and human produced products and services? And will the human products and services be have to
be incredibly cheap to compete with the AI ones? >> I think that they're going to be reversed actually. I think that what you're going to find is that the human services and that human stamp of authenticity is going to command a price premium and it will be more experiential. It'll be kind of like like mass-produced clothing where like you know you have a t-shirt from Hannes or Fruit of the Loom or whatever. Uh or even just like you know button-up shirts like if you go to like a Macy's or J C Penney and get something off the shelf it's like it gets the job done. It's mass-produced. Um, but that's not really what I'm after. What you will get is the people who are buying for status and they'll those
are the kind of people that do things like these shoes that I'm wearing right now handcrafted by uh Giuliani Armani in Italy, you know, like with his hands, you know, that's the kind of status and buying power that I have. And I feel like we've seen this over and over again. It's the same. We see it in hospitality where like you know you have fast food like McDonald's that is >> delicious, >> very in also very inexpensive. But then you have things like, you know, like you you go to like a cowboy star or um a Ruth's Chris and it's like I don't know that was somebody made that today or like even more artal stuff like this restaurant that we're going to is on the farm where they raised the pig that
we're about to have the pork belly of. So I think that what you'll see is this might be an interesting opportunity for blockchain to make a resurgence where it's blockchain verified that this was made by a human as opposed to AI like u and that will command price premium at least >> 26 in my double down on blockchain. Well, the thing the thing the thing that's going to h the thing that has to happen though is that you have to be really really good. Like it's going the there there will not be a like there's not going to be a >> no it's just it's not going to happen. You have to be really freaking good. We kind of see this with >> I feel like the same thing happened with music too where act and the thing that
happens is that people actually crave a little bit of imperfection in the the rough edges that they >> Yeah. rough edges. Um, and we see this with things like it's like the thing that makes a uh like a vinyl, an old vinyl so powerful to listen to is because it's actually not overproduced. Um, that there is that human imperfection that comes through. And you even see this with things like a live performance. You know, this is why live performances are so valuable. It's because there's chaos and there it's not programmed. That that's why podcasts are so popular. It's because it's unscripted. Um, and I think that we still long for that. It's just going to be it's not going to be possible to have a middle ground
anymore. It's either going to be high-end, high touch, high involvement, or low touch, good but not the same, like just a difference in experience. >> Mhm. Yeah. Uh two thoughts on that as a follow-up is one I think that anytime you see the cost of doing something go down where it becomes a lot more accessible then there becomes like a new bar where you have to go you have to make it even more high quality and obviously kind of handmade humanmade in order for people to see that difference and appreciate that difference. So yeah, like now you know maybe the cost of like getting an influencer has gone down or maybe before you could only get an influencer who was a Hollywood movie star and you had to
give them a million dollars and 10% royalty in all sales. Now you can seed your product to a hundred Instagram influencers, give up no equity, and you can do it for, you know, uh, $100,000. And now with AI, maybe you can make your own AI influencer andor you can just create ads of uh, people that look like influencers with your products that you can use to advertise for basically free. But now once you've like used that up, once you've exploited that as a tactic, and now you want to go the extra mile and you want to go beyond what that can offer you, now you need to do things that AI can't do or that's obviously not AI. So like, yeah, every every pose that you think that an AI model can do really
well, now you're going to go the extra mile to find a model who doesn't look like anyone who's AI. Maybe the model looks a little bit imperfect. Maybe the lighting isn't perfect. Maybe they're doing something or they're making expression or they're holding it a certain way where it just people can see right away that that's not AI, but it's also more eye-catching than what would have been AI cuz that's the key, right? It has to just it can't just be different. It has to stand out and be more eye-catching. Um, and and and I'm using example of like an AI model because it's a really easy one to visualize. But again, this goes for blog posts, this goes for tweets, this goes for >> um landing pages, this goes for, you
know, software UI abstractions, so on and forth. It applies really to anything. I wonder if there's a world where it's like because in my brain there's you could be like, "Okay, the law says everything needs a watermark." If it's got AI, you know, you got to say that it's AI or whatever. I I don't want to do that. >> Instagram. >> It's a little weird. I I in my in my head I think something there's like a reverse where anything could be AI but if like the AI has this stuff which you can't have in it like it it doesn't have this you know like there's a badge or something that like there's not this isn't in the AI and so if I have it in my video you know that this is a real human being or something like that. But
I just I don't know to the little watermark that's like AI was involved in making this. >> But instead it's like oh there's a little badge you can put on your thing that's like that it's not in the models. It can't be in the models. It's like legally not allowed. So if you have that that's like a verification like because that's the problem like the government has to like put out statements or like the president has to say something >> and the models are so good you could like make whatever you want. How do we know it's actually him saying it? Anyway, we're way down away from vibe marketing. >> No. >> Yeah. Well, I want to get to two other things based off of that really quickly. Sorry, there's a fly buzzing around me
cuz my door is open for my dog. But, um, >> it's not real. >> Yeah. Uh the the other part of this things that I think is really interesting based on Zach's point about uh commanding a price premium is that I think on the product side of things, what I think that we're we're going to see happen, especially with a lot more software and services is a lot more personalization because if we can't apply the AI to the marketing, we can probably apply the AI to the product more. And I was thinking recently about our client Budible. I'll call them out. Give them a little shout out because what's really interesting about them is that they they're sort of positioned as like look we're the internal knowledge base for X Y andZ
industries. So your employees can go in and ask any question and then get an answer within seconds instead of having to file a report with HR and maybe have to spend weeks on something that's not even a good answer. And um and what's really interesting is that they have each one of their clients upload a unique corpus of data, unique training set, and then like for them, the product is tailored and personalized for that client and all those answers. And I think we'll start to see that a lot more where it's like instead of just logging into your I don't know your HubSpot dashboard and you get your CRM with all the different tabs and buttons and things. It's like no, this is the conversion factory HubSpot dashboard
with all of our only the buttons that we need and a bunch of new buttons and new interfaces that are personalized just for us. And now it feels like a much more customized experience for us because at the end of the day, people are are going to pay for things. They're going to pay extra for things that they know the AI can't do. Unfortunately for software, there's nothing that AI can't do. >> There's no like, you know, handwoven or handcarved uh equivalent in in software. You know, you can say this this shirt, you know, the the stitching was specifically handwoven by, you know, a craftsman and no sewing machine in the world can do that. It's like, cool. I'll pay a premium for that. There's nothing that a
programmer would write in code that gets compiled down to binary and a computer chip that an AI can't do. But what can an AI do that that we can't do? It can customize and personalize the experience of that software infinitely, right? Um, so I think that we'll start to see that a lot lot more where it's not just a off-the-shelf experience for every software that you buy that it's a personalized experience. There's a lot more customization involved. Um the second thing here really back to the to the vibe marketing theme is that beyond just creating marketing assets cheaper, faster, better, I think the real magic is actually in the workflows um and this more like agent experience where for example I can set up an
automation to scrape my competitor's Facebook ads, feed that into GP APT40 or whatever image model is the best at that time. Have it create personalized equivalents of those ad formats for my brand. Go and then write the copy based on the creative, upload into these Google Drive folders, grab those and upload those into Facebook ads and then like send to me for review. >> Or so on. Like there's so many different workflows that you can't even imagine that are like, >> you know, we've talked about this. we've been looking into like Gum Loop and Lindy and it's like oh what if after the sales call it automatically grabs a transcript, generates a follow-up email for us, sends them a recap of the conversation,
um connects with them on LinkedIn, sends them this personalized message, follow-ups in a week if there's no customer record detected in Stripe. Like there's so many cool automations where the workflow could be like 20 steps long and it's as if someone were doing that on the back end. You know, it seems to me that what the the key thing with like the vibe coding and the vibe anything is it's it's more about a um just getting rid of minutia. That's what that's what this is. It's you you cut out those interstitial steps and get rid of the minutia that is needed to get to your milestones before it can actually be something that is worth showing to the world.
>> Mhm. Like because what you just mentioned there and something that I've been looking at with Lindy, I'm like, you know, checking out Lindy AI's notion integrations with Stripe and all this other stuff to see like what what are those small little steps that you have to take to get to a specific uh way marker before you have to do something else. And the more of those you can cut out, the better. Like any part of minutia before it's like milestone, great, I get to check in. You're essentially becoming like when we when I was doing forance graphic design before we started conversion factory, you'd have those milestone check-ins with with clients of like, you know, hey, we're going to check in on progress here. Our
expectation is to have this and we want to review these these elements. Now, you don't need to have as or your milestone check-ins become far more frequent where you're able to get to those stopping points refilled. This is actually something really good and I can see fantastic progress um with with this stage and we have other stages to go but I'm cutting out all of that other energy that is needed in between those uh core steps. >> Yeah. >> Yeah. There's like I I Okay, as for all the anti- AI propaganda that I spew um I use it real like I do use it. Um, but I I I totally agree. I think Corey, what you said about like the the change in the workflow, not necessarily like the work that's being done like it has to be
done this way, but it's like we're cutting out some of the BS and we're changing our workflow so that it's just fa faster and more efficient. Like I today I haven't had an email from one of our past clients just asking like, "Hey, like I'm like still figuring out this SEO thing. like can you kind of give me some tips like pointers things like that on their current site? Um, and I was like, I could spend a lot of time thinking about this or I can think about it somewhat and I can also have AI think about it and feed it the URL and send like copy the email that that they sent me and be like, "Hey, this is what the client said, uh, past client and this is, you know, what their URL is." Like, what do you think? Like, tell me what
you would do. And I'm validating what they're saying. And and then I'm like, "Okay, cool." Instead of this being like an hour and a half of me spending time on a past client, I have like, okay, I got everything I need and a good frame of reference in like 10 15 minutes of which is mostly just the time in my head thinking. It only took the AI like 30 seconds and then I can send him a loom and you know give him a breakdown. Um, so I I totally agree. I think it's the workflow and it's cutting out the BS that is the true value that we see right now. The question is how long is it going to be before it's more than that? And it's like oh it does all the work. >> It does basically everything. >> I think that that as well as taste like
at what point does it actually >> have taste >> come up with ideas on its own in a really provocative way. >> But those to be provocative. >> Well, I was going to say those might those might be a podcast in of themselves. >> Yeah. Yeah. short term though. >> Sorry. >> No, you >> you >> Okay. I was just going to say we should have a whole episode of us trying like we all independently go try to get AI to do or say something ridiculous and we bring our results to the end. Okay. So, what I said doesn't matter. Okay, Cory, you go. >> Okay. The next one off topic. I was just going to say speaking of minutia, one of the things that I've noticed myself doing way way more is just speaking and dictating versus typing.
>> Mhm. >> Um >> even the most simple things. Uh >> whisper flow >> or GPT. Uh well both because like chatbt on your desktop app has a really good um like voice to text feature um where it basically uses their own um speech recognition and speech transcription uh API. >> Superflow and Whisperflow both use that same service essentially. But yeah, I've been doing that even with this book that I'm working on. you know, I'll I'll see the feedback that people are writing about certain sentences or sections and then I will essentially dictate what they say line by line and then I'll feed that back into cursor because it's has my whole chapter in there and then start editing that way. Like I'm not actually
in there typing words or making changes on my keyboard. I am purely just dictating and I've been doing that for, you know, the copy that we're working on for clients. I've been doing that for emails that I write. I've been doing that for for tweets. I've been doing that for um even just like going through uh like I did like an onboarding audit for a client recently. So I just signed up and I spoke out loud what I was seeing on Loom and then I fed the transcript from Loom into CHPT and I was like, "Okay, now tell me all the what I said like structure that into an onboarding audit essentially trying to answer these questions and format it this way and an outline." >> Dude, why why haven't you told us that
you did that? Cuz that's a great work, >> dude. I've just been vibing, dude. Sorry. is viable. >> Well, I've been because I've been using Whisper Flow as well. Um, just speaking is so much more effective like because I can it's it's thinking out loud and that's faster the way that Yeah, cuz me having to type things it's it's very curated and it's cold. >> Um, but you know, I can do all that after I just uh brain dump everything that I'm seeing on there. And I hadn't thought of having GPT then do that culling and refining for me. >> Well, yeah. But it's great. The I think the ideal scenario, >> this is, you know, future Zach talking is we have the screens on contact lenses and then we're able to just walk and we
can have we're able to dictate just using our voice and it's like, "Yeah, move that there." So that way I can be walking um out in my neighborhood and giving feedback to clients verbally and looking at it. Yeah, that's the dream, Nick. >> That's a Black Mirror episode right there. It's >> some work life balance, my friend. Heck no. >> Um Corey, I was going to ask, are you meaner to the AI when you type or when you speak? >> No, I'm super nice in both. I'm sure you are. Sure. >> I'm >> I say please. >> Yeah, I do too. >> Please. Please po please and flank you. >> Very rarely. Only if it's like majorly effing up and it's not understanding what I'm saying while I like all caps, you know, rage. Uh
>> like no this not that idiot. >> They're just like tearface emoji like sorry Cy. or well I think I mean a a good example of this is like when we were discussing the um frick I I see Has C corps and S corps >> and I thought I was like man I don't have time to research like what the pros and cons are of this like GPT please help me out um and that's a perfect minutia example of I have no idea what this is I just have a question and a theory that I would like tested. I need to start using it more often when I am walking with the uh um the voice like the actual >> Yeah, there's a conversational version.
Get to talk with Veil from OpenAI and say, "Vale, I've got some problems I need to talk to you about today." She's >> like, "What's wrong, bud?" >> Oh, wow. Sounds like something's on your mind. How can I help you? >> I I don't want to talk to them. That's the the one thing I haven't tried yet is talking to it. I don't want to do it. >> I actually don't like it. I just purely like the >> the the speech to dictation feature. Um the conversational to me like >> takes me out of that kind of vibing state of mind where I'm just >> thinking out loud and I'm more like trying to curate what I'm saying, >> especially in like a a short amount of time because then it's like what did you say? I missed I missed that. were like,
"Tell me again." Um, but yeah, that the definitely like speech is definitely a big part of Vibe coding and it's a big part of Vibe marketing now, especially for content creation, you know? I mean, even just months ago, I was like, "Dude, why are you even trying to write a blog post? Just talk and just have the AI start to organize your thoughts." That's how I finished the last like >> third of the book was I just did all that speech dictation into GIPPT and then kept copy and pasting the contents because I was like this is too much for me to like write and self-edit and like yeah there's there's definitely an art to writing and writing is good cuz it makes you think very clearly but when you're just trying to get pen to paper
and like thoughts out of your brain >> there's like nothing better right now than the AI speech to dictation. >> I'm in that. >> All right, we have been talking for a long time. It's crazy that on an episode with no notes, we talked for this long >> and we could probably talk for another two hours. Uh, any final thoughts before we close it out. >> Uh, if anyone has a very specific Vibe coding or Vibe marketing use case or workflow, please let us know. >> Yeah, show us what we don't see. >> Yeah, I'd love that. Cool gentlemen, great pod. Thank you for teaching us Corey about Vibe marketing. Next week we'll be talking about Vibe something else. Who knows? >> Tune in next week. >> Okay. We will see you next time. Bye
>> bye. [Music]
Vibe marketing, agents, and which of these tools have earned a place in the stack rather than a place in the newsletter.