AI Slop is $100M Business in Plain Sight

good morning my name is Parker Rex and I led product for a startup that did $50 million a year before we exited after that I tried to do a startup with a couple really really silly ideas and hindsight but quickly came to realize that I needed to be a developer not just a product manager who was technical so I've spent the last three years just doing full-time AI development and some good client projects as well and on this channel basically just building it public I'm trying to scale the AI services that I do to 100K a month and I do a daily update so partially to keep myself accountable but as I keep posting these people will find out about them and'll be able to see the stuff that I'm learning all along so the schedule for today is got a couple things that I want to cover but first off I think the mic will be better off I Know audio in the last one was pretty rough but we're early here so probably won't have many people watching it anyways which is fine but first I want to do some video planning I have a new system in place for this is really just like taking what I've learned in product management and applying it to video editing or video planning I want to go through some of the benefits of roles and prompting and everyone kind of knows about this stuff where it's like oh you know act as an accountant or you are or a writing assistant or or whatever there's like some basic ones but I think that you can get way more in depth so it's like oh you were an accountant or you are a blank like the whole point of that is like you are a x but I've found that you can get even more detailed with some of the business planning stuff that I'm doing so I wanted to go through that quick client lesson um this plays into the overall cohort type of working with less technical people versus Technical and some thoughts on that we're going to go through the website that I built out kind of talk about the technical side of that what I'm planning on doing it doing with it I want to talk about a trend that's going on in SAS where superbase is like blowing up but I also see a problem which could be an opportunity for business owners like myself um and then going through some website stuff and then I want to actually get really technical towards the end on some code which will be fine because it's multi- aent and I need to get that project running so first of all on the brand side and the video planning just make sure everything's good here my levels not dropping any frames cool yeah seems fine it's daylight savings here in the US I think it's daylight savings today daylight savings yeah so it starts today which means that we skip it a month or sorry an hour forward which is really quite weird not sure why we do that but that's today so cheers for that a little coffee let's see so on the video editing side I first want to to see I talked about this in the video yesterday but um my YouTube's been flat for pretty much ever and this is the guy boom I've been watching this guy's videos they're good he's he's taught me a lot about No code stuff that I did not know okay so it looks like I'm weeding out a lot of the people that I've had because I had a small audience I think I GW this to like 5,400 or no sorry like 5,000 like quite a while ago so right now it's just like hovering around the same number which I'm fine with and I kind of expect um as I'm just reactivating this channel so if you think of a YouTube channel is an asset it's kind of what it is it's an audience it's one of the harder platforms to to crack you can build an audience that may not resonate with the content that you make later so I've had this channel for a long time and you can see I'm G to pull up my lifetime stats but it's taking a while because it's like wow you've had this thing for a while so I had it for Wow since 2013 but I didn't start posting until see 2020 so let me go to 2020 to date so let's do when was I doing Feb of 2020 who no December of 2020 like five years ago I started posting on this channel so let's do that to now cool let's see but you'll noticed that a lot of the videos I was making were in the like product management specific Arena so it looks like I posted like day in the life of a software PM um building apps with notion talking about our idea to sale talking about all birds like kind of all over the place how to win in product management like basically really nailing down the product management keyword and then after that you know it's like kind of this mixture of like couple different topics be product management it' be Day in the Life it'd be a review of some piece of software but those are really different things like I think the Vlog style ones are just such a pain in the butt to make and don't really reflect what I actually do every day so they're too hard to make even when I had a day job but you can see that you know you start to pick up some steam you start to build a little bit of a subscriber base but then recently you know I basically stop posting for w stop posting for three years you come back and it's like you're almost starting from zero because you are now posting videos that are very technical they're about AI so I've just noticed that like where I you know I did maybe say 24 let's call it like 30ish videos and in the 30 videos there's also a gap in between so now I'm just getting back in the kind of rhythm of reigniting a previous audience slash kind of moving them out but I'm not really that focused on what's going on every day or every month because I'm just getting back into it I know that this is an asset and it's like investing um s investing like when Snapchat IPO came out and like you just learn it's like compound interest but with audience it's like get in the habit it's like your dollar cost averaging which means in investing world it's like you put $5 in every single day um I know if I do you know three videos a week on the main Channel and then a video every day on here that I will naturally get better at speaking on the camera naturally uh basically sharpen up those muscles that have maybe gotten a little a little flabby it's a tough visual but then when it comes to the actual video ideas this is a really popular framework in product management where it's like if you had a feature it'd be like feature name it'd be impact confidence ease and then you'd have ice and so if you had a feature and let's say You're Building out a calorie tracking iPhone app and you want to say that you need AI Vision feature where you take a photo of the food and then it gives you the uh calories and it would give you a rough estimate of the macros like the proteins the fats and carbs you would guess you know you'd have an overall goal which is like increase uh daily minutes used so if that's your goal then you would base these scores one to 10 for each of these you'd have this list of features and You' go through it and you'd say Okay um what's the level of impact that I think having this AI Vision feature will have on a calorie tracking app that a building well I I think that this is probably going to be like a nine cuz a lot of people kind of expect that out of that sort of app and then my confidence in the impact like I might be thinking that this is going to be so high impact but am I that confident in it like it could be really good but will it so it's this additional kind of feel test where it's like okay if it hits it hits but I don't think I'm not sure if it's going to hit so in this case I think yeah it's probably nine and then e well okay based on the implementation are we going to be storing the photos are we going to be uh the ephemeral are we going to be adding them to our kind of library to be able to go through them later to train it on a bigger model all those things would be in your head and then You' be like well out of 10 10 being the easiest where is this at you say okay well I think it's probably a five because I do want to train and so then I would do that and then i' just get the average so you're trying to get a higher ice score and so I kind of am bringing this into here it's like well what's the ease of it to make and then How likely do I think it's going to blow up and I just call it an e score so what I've started doing is just that so I'm G to kind of go through these real quick but I think that this will be a good barometer for me because I start by going and I look around at what stuff is working and I talked about this in a previous video but the way I'm thinking about this is I have two kind of niches that I'm trying to Target with my old channel so on my channel let's call it um Main and there's the channel I have two so the first one would be uh kind of like Mass audience AI content and then that scratches the itch for a lot more people right because it would be stuff that like my mom or my fiance could watch if they were interested where a topic might be like chat GPT in five minutes or in three minutes and that's based off of me going and typing in on YouTube Just classic keywords that people would want to see because this space is so early right like it's not it's not old at all so I might say chat GPT tutorial and then I can look and I can see like I know I made a video that basically mimicked this same setup this was a master class style 27 minutes but you can see that I have an outlier score um kind of see like this one blew up but these would be kind of bigger um audiences so it's like you know chat BT tasks and three minutes this sort of stuff now what might be interesting is actually like this would be probably a good video where you have this guy that everyone in the tech bubble knows about but people outside the tech bubble like if I told someone on Twitter do you know Andre carpy like yeah of course but if I said that name to someone that's not Twitter they'd be like what who but maybe you could reframe this and it' be top five lessons from the guy the creator of Chad GPT and then I drop a note in there and I'd say take Andre karpathy video and extract the Five Lessons that apply to the largest audience and you know what since we're here we'll get back to the second Niche but this is just a really quick way that you can do this is like I'd open up notebook LM this would be how I'd use AI to make a video about AI right kind of meta but I can just go to YouTube link and then paste in the YouTube link and because I have notebook LM plus I can add a bunch of stuff in here but it's going to go ahead and just train on all that you can notice like that took 3 seconds and I can make this podcast about it but instead I want to take a prompt that I have I think I want to take a prompt yeah you know what I have a bunch of prompts that you can check out that might be helpful for this like if I wanted to make a blog post I'd go into my Air table which you can access um I I don't think I use anything on paid when but if you go to par.com and you go to resources AI tools you'll see this database in here and under YouTube to blog post like what I could do here is I could take the notebook which is basically just taken in that two hour long video and it's summarized it so I can query against it I can write stuff to it and it'll give me answers and if I wanted a blog post then I'd grab the one from YouTube to blog post and if I could if I wanted to I could pick any one of these so I could say okay what type of blog post do I want do I want a guidance tutorial comprehensive I just want to just as an example let's try to see how to transform technology review to a Blog so I'll grab all the stuff that's in here this is written inside of a code block was originally typescript file but in this case we don't want that obviously so I have this right I have the source and then I'm just going to paste this in and let's see why is it not limiting me is there a restriction what I'm guessing is you can't have this long of a prompt which usually you'd see something that would let you know that see generated comprehensive let's take this and we're actually going to use AI to reduce the size of this uh cut this prompt and half but keep the semantic meaning and let's fire that off because I'm guessing it's it's a length thing like let's see yeah it is okay so let's see if it's cut in half how it'll do maybe it might might need to be cut into a third let's see this to uh yeah cool so let's see how this does in terms of the blog and then I just wanted to give this as an example how you could kind of approach um taking a YouTube video and literally just turning it into a vlog and so voila so let's see the output of it Andre Okay in his latest video Renown Searcher provides comprehensive insightful look into the Practical applications of large language models like chat GPT moving beyond the underlying fundamentals discussed previous video he delves so Del this is a meme um you would not want to be falling under this Del polyram founder of Y combinator if you've heard of it um he actually made a huge deal about like oh you should never use the word Del yeah boom Ed the word Del 17.4 so it's a sign that's written by chat GPT when you use the word delve so you would you know maybe edit that prompt a bit but for the sake of this this would be a nice first draft and then I can kind of go through it and tell it what not to I'm actually curious if there's a way that I can have a system prompt for this is there a way to have a system prompt for Notebook lm's product by Google let's see yes you can Divine a system prompt okay so it does not have a native one there are work arounds so you'd have to explicitly put that in interesting okay well for the case of this I would probably just take [Music] um this chat I'm wondering if I how I could keep the previous chat I haven't used this a bunch as you can see but I know that it's it's kind of the go-to oh here we go conversational style custom boom got it okay don't use words that uh people Beyond eth grade wouldn't know for example the word delve try to write in a conversational tone 75% Spartans 25% um conversational bar tone cool so the eighth grade writing thing I think is really good and then I could just refire that same prompt but in this case I would probably just trash this so I I think it is weird that you can't out a new chat but give me a YouTube outline for a video called five most important points about chat GPT from its creator and even this I would take this and I'd go back in here and I would say I going to get better title generate 10 different titles that are increasingly more click baity for a video that breaks down the five most important points or takeaways from Andre carpy's video on llms he's a creator of chat gbt and this needs to be for a mass audience example cool so that should fire back five mind-blowing AI secrets from a chat GPT Creator chat GPT Creator just revealed five AI truths that change everything what weird as these work you know when I read these I'm like H so lame as a Creator but then I find myself like clicking on these types of things so let's put this in let's see what it comes back with and this will kind of get the ball rolling for me now I'm like okay like when I come back into the video ideas section um here we go so LMS are like internet knowledge zip files that's good LMS have a limited working memory LMS think like humans LMS need tools and customize your AI assistant cool so what I do is I can save this to a note and then if I wanted to I could come back to it so this is kind of their way of like letting you out of the chat I'm guessing so I just take this and then for the future video this will actually I kind of like this one I think this would probably blow up um so I'll just stick that in there for now and I'll probably add a column that is notebook LM notes notes notebook LM just so I remember that the notebook thing is probably a good way to to be doing this because it does seem to be pretty helpful so I want to put this one in um as needs to film cool and this will go on the main Channel and then the series this will be like a listicle style and I think this would be really easy to make and I think it has a really good chance of blowing up so now I know that I should probably do this one and then I need a hook for that that so I'd put this in there I wrap this overflow no make that smaller oh so now that I have that there this should be kind of how I go about it right I think that this would be a good video to do I think that people would read that they'd be like okay well that makes sense so that brings us to this other Niche which would be more like on the coding side of things and I I do see this as also being really popular I don't want to split the channels but this would be like you know not not that Niche but way smaller of an audience where it would be like AI content oops still figuring out this freaking pen AR content for developers I talk about this a lot but I think there's this kind of schism where like maybe like 50% of people think that AI kills Dev and then I think the other 50% think that AI grows Dev and I think this is I lean more on the side that this is a bigger piece of the pie than that one I think that like basically the ease of use on AI development tools will just flood the flood the Zone like if you think about it um like even this camera that I'm on like this is a a little thousand camera that made it a lot easier for me to like shoot in 4k before like you know it's 2025 now but in the camera use case a 4k camera 10 years ago I wouldn't probably have one that's at $1,000 doll right so it's just easier same thing with uh you know in 2008 you get the iPhone and it has like the ability to take pictures now everyone's photographer this thing is literally just a camera and then you get much higher quality photos you have a lot of photographers who were like oh it's just an iPhone photo but guess what a lot of great photos are taking on it now and that's because the tools are better you have same thing with video editing tools where if you had Premiere Pro 10 years ago then they probably hated all the people that were making Tik Tok videos as they flooded the zone so that's kind of this sort of content and the benefit to the audience is like there's so many people that are now like Vibe coding and I think that you can kind of like ride this wave of making content about this but then like the counter signal of that is like what's called slop and I've noticed this and it kind of brings me into another topic that I wanted to talk about but we can we can cover that now where it's like this idea of the superbas slop generated code now this is a really popular topic where if people are basically like they're building an app and they want to get it you know from call zero to one which is very different from what you call one to 1. n which should be like an improvement so this is going from like nothing to something versus something to something new and a lot of people are able to get like the first version of this by basically prompting a million times and it's good like I think it's a a net positive uh for all these developers that are new to come in and and figure it out and they basically just do it like they oops they jump into curser or wind surf or whatever and they oops they open up their app and they have the series of like prompts that like you know their code bases in here and they have this button here that's like the submit button and they're just like basically spamming this the whole time and it could be done in the ID or it can be done in the CLI but they're just you know kind of like gambling where they're like rolling the dice that this thing is gonna you know nail it and it's kind of like gambling where if you've seen the like slot machine meme like this is uh like pretty much it old person like this is cursor this is this is exactly it we're like people devs are literally just doing that thing over and over and over again and that's not good um I personally know this because whether it's on client projects or even my own projects you're not gaining the knowledge of how the stuff works so you're not going to figure it out you won't know how to undo the swop I don't care how good the models are going to get like you should you should know how the the fundamentals of it work um back to like the camera analogy with iPhones and it getting easier and easier and easier like there's still competency to knowing how to use the the new tools so someone that's really good with the iPhone probably knows by now like what brightness and exposure and contrast and sharpness and vignette what those things do um which would be the same thing like if I'm in cursor this project this website that I I have one that uh you know it's my personal site I was was able to Vibe code this by just like shooting stuff in here and saying okay like we need this page for the uh Google Knowledge Graph we need this page for a biography put these in here all this stuff but the second that it got to um something a little more complex it's literally just not like a static site but it's something like okay we need an admin dashboard okay we have the admin dashboard it can make the login page and I know like there's ways to inform the model to make it better with an mCP server which I did but the second that I had to bring in like a new library and have super base working with the off and oh what type of off do you use it ran into issues where it's like oh like it's putting in this redirected thing it's showing that I'm logged in but then when I click this it's you know it's broken right so then it gets in this Loop where you have a a bunch of different migrations where it's like trying to figure it out and you can see like this is the basically the product of vibe coding is a bunch of these and now you know we got like what 15 different SQL migration files in here and had I just done this by hand I wouldn't have been where I am now which is where it's broken I would have just done it myself um and I would have used AI but I think that that's the difference is like because I know how to do this because I have done this so many times without just spamming code but I kind of wanted to do this as an experiment to see how good the models are you know I would have saved a lot of time and even like trying to get it to add tests and stuff like total total um runaround instead of maybe like I think the point I'm trying to make is like the people who are kind of in this Camp of like oh Just Vibe code everything I don't think that's good I think you need to still know the fundamentals you know like Michael Jordan isn't as good as he is because he like you know he has the fundamentals down and I think you can be the Michael Jordan of vibe coding if you just knew like you know fundamentals of the following like this would be like the curriculum like if I restarted like I started coding for real three years ago I've been on and off coding since I was 18 but like didn't take it serious but you'd still want to know like okay you need to know like HTML right you need to know CSS you need to know JS you need to know TS you need to know react you need to know framework like next because if you go and even something as small as like um well how's it doing it in is it in react 18 is it in 19 because there's very different different patterns there and the models are always going to be a little bit behind what's going on and I think you could probably air on the side of just having it go the way that it wants to go but then when you come to debug it it's just going to be a mess so I think like even like it goes for like the next 12 Pages versus you know now we're at next 15 to up and a lot of people will then say Oh okay like I'll mitigate this risk by having a you know cursor rule and you can do something like this where you have a skill set context prompt where I think this is a good thing to have like it's kind of a cope but you would say like in XML like here's capabilities here's the languages here's the Frameworks here's the links here's all the stuff that I use and but then it's just an extra step so I think you know this will kind of be an area where a lot of people will jump to oh we don't need to pay a developer because we have this guy who kind of knows how to Vibe code and then when they build it and then it's a mess they're confused and you see this on upwork actually because I've been doing contracted work to cover cogs for my own sass which we'll run in a little bit here but if you look in here like if I typed in next right like you'll just see so many many of these and a lot of them I noticed just from kind of breezing through them is it's a lot of code bases that are just garbage dumps and it's because it's like they went into you know like lovable or or something like bolt and those things just they get you a first version that's like kind of like a bar Treck where it's like oh it's here but then you know that the you know Foundation of your projects maybe not that great um maybe it's gotten better I don't want to hate on them because I that was just my first impression but I think that this will be a a huge market for people that were like kind of like I considered myself like a mid-level engineer like I three years but then I did double time so I want to call it like forish because I just work a lot but I think that this will be like a huge Market um and that just plays back into like I want to help kind of get those people who are in that like Junior phase to get to know stuff Beyond just like here's this prompt it's like I'll get people in the door by doing videos like that like I have uh don't really the only video that I've had any sort of like viewership on recently was a video like that where it's like um see I have to scroll to the third page to find anything that has like good views on it or maybe not hang on it was a cursor video but I know how split testing yeah so it's it's this one yeah so 11,000 views and 403 likes and that's just for showing people like what my setup was for maximum productivity and you can see that like people are obviously obsessed with that because they're trying to figure out like how do I get this thing to go so that'll be really like the the way that I service that audience and then they'll be like oh this guy's legit maybe I can learn some more stuff um yeah so the next topic I want to cover is we covered this I do a strike through here I'll just do it by hand we did video planning we did this want to talk about yeah let's do this benefits of roles and prompting real quick because I know this is already 30 six minutes and I got to get going on work today but um one thing and I spoke about this in the video yesterday but cares is this kind of benefits of prompting where oh this is an interesting prompt if you do want to find video titles for something you already recorded like I did this for the last two daily uploads but you can take the transcription that you get and then prompt it to give you a title that works and it actually it's pretty pretty legit and the way that I get those uh transcript is through this product called aonic that I got from again that make uh Creator so where was that prom here it is so one thing that I think a lot of people um mislabel or forget when they hear the word Creator or influencers is like people that are absolutely crushing you you kind of like get past the the name influencer after a certain point like when we think of influencer we think of like girls that are uh like shaking and bikinis and they get a lot of followers and they sell you some like supplement but there's people beyond that right like Joe Rogan for instance like he's been doing podcast since 2009 and I'm curious like what is it stuff like what stuff does he say that you know helped him get in the Cadence of continuing to do it so that he could be the largest podcast in the world and then another one would be this guy Luke Belmar because he talks a lot about community building and I think that's something that I'm well I know it's it's something I'm really interested in because I'm building out a community called trouble-free which basically lets people join we have six members now but it lets people join to talk about AI talk about how they're using it so it's going to be servicing two different niches it has all my stuff in here I do need to make automations that keep this up to date because there's way more stuff than what CCI here here but he talks about um that kind of like community and how defensible it is against AI which I think is really smart but this really takes like a prompt about the context of the business that I'm building so has this this big thing about the background of what I'm doing what I plan how I want to be like the Accenture of AI services but from someone that's actually in the field doing their things every day and then um about like the marketing funnel stuff that I was talking about before with like the two different imges and then how do I bring in the research and kind of playbooks of these people that I want to emulate and then apply it to what I'm doing so I can say you know I'm doing research on community building and content generation and I want to get some research on how Joe kept up with content and his story and then you know some details about like maybe there's Snippets and then for Luke like what's his approach on both community building and content creation how did he grow it what's his Playbook so one that's like older because 2009 you know 16 years ago Luke bmar I think he launched his stuff in like 2019 so much more recent you know 10 years more recent so I give it all this and then let's just see what it getes back so Joe R's podcast the Jo experience began 2009 signing a deal with XM so two it looks like yeah so he started by getting on the schedule so we know that we need to stick to his schedule here whether or not I have zero viewers on this channel uh for a year or not it's so good for me to be able to do this because it's basically like I do this anyways like this is like a brain dump but I'm just talking I'm getting the skills I'm strengthening those muscles around talking to the camera so I can use my hands keep you in the box and all these things um so that eventually in like six months someone that's watching this will see a much better product and then these won't get views for a long time right but I know that when I am successful and things are really popping that people would be like whoa this guy like called a shot and it's because I'm going to be following the schedule and he's on two episodes a week couple Milestones it looks like he flashlight I don't even know what that is that like what I think it is but like two years in Serious XM deal kind of crazy four years in starts putting it on YouTube 2020 exclusivity $200 million contract wow okay so how do you be the Joe r of tech right now my coffee is cold so stuff that I would take out of this is like obviously stick to the schedule expect two years of talking to the void which is for me like that's the contract that I'm signing it's like okay it's if in two years from now this is working would I regret doing it for two years talking to nobody probably not um Community Building known for his entrepreneurial Ventures started building his audience through content creation launched in 2019 focus on crypto and personal development okay so those are two giant niches right and so so my two niches would he went Ecom again huge one right so he's going like wide wide wide so M would be Ai and actual development maybe later I go into something else but for now it's AI like broad AI usage and then software development for AI authentic content paying members 200k a month cool those are good number survey note the section provides comprehensive examination so oh Milestones 2019 for sponsor this could inspire you for your proposals yeah I guess so formerly named The Joe Rogan Experience what is it called now I'm pretty sure that it is what it was called I don't know expanding reach Global scaling via upwork yes digital platforms start uploading videos under the account powerful Jerry okay so he did start doing clips got it so you did eclips channel four years then mve to Austin yeah so I would change this a little bit because it's like it's taking like what I wanted out of this was it would give me um strategy but what it's doing is it's just like tying the similarities okay distribution playlist or Playbook early episodes suck I didn't know what I was doing I didn't think anyone was listening it was just for fun there's a lot of us just hanging out with Comics okay yeah and I do think that this part about bu building relatable brand is important because it's you know it's not like I'm doing 100K a month so I'm not here like standing in front of a Lamborghini it's like you know I went from high-paying job to grinding startup to trying to raise money didn't work to now saying okay I need to do AI Services did 20K the first month I didn't uh go like the VC route bad quality of Revenue wasn't repeatable wanted higher quality Revenue now doing about half of that and like per month but it's higher quality Revenue where it's repeatable and putting systems in place so I think it is like going to be way more relatable for people because raising venture capital is like such a small percentage of the world does that it's like such a little like Club H incretion so yeah I could take these two videos and then I'm actually going to pump those into see why did Joe Rogan start his podcast no why did Joe Rogan start his podcast and I just want to pump these into what him start J I can pump this into notebook LM and then just clean some insights from it right so I would just call this how I use LS video and then I'll go back and I want to create a new one and this will just be for popular or let's call this a Joe in Playbook and then I can take a bunch of those different videos what are the main points about Joe Rogan's Playbook that I can um use to gain insights into my strategy cool so I can come back to that later yeah I don't need to read all this to you but I think this this is one way that a lot of people are not um using like the role playing stuff it's like apply role playing to your own plan um what I am going to change about this though because I don't love the output I think it is definitely helpful but want to put instead of tying my current strategy to things that they're doing take the broad Strokes of the strategy and man ulate it to match ideas that you see from these two creators get creative come up with new strategies it's very very important that you get this right so that um I want to give it Stakes so that I don't disappoint my future family this is actually a paper that came out um there was a paper that came out that basically said if you give it the emotion like if you're if you're psychological about it if you if you give it emotional stuff then you get like 20% better outputs which is crazy like literally putting in I mean it can go so deep where you the founder or you're a responsible for giving the information and answers to the person asking the you know the indigenous natives of this Forest are going to have their Forest chopped down if you don't do it it's so important like please it'll give you better stuff if you say extreme stuff like that but yeah that that like covers the point on the the role playing um just quick thing on like client work that you know I wanted to bring up is like sometimes like depending on what type of client it is like if you have like uh design oriented clients if you have Tech oriented clients if you have uh operations oriented clients like you're gonna have different expectations from them and so for instance like a tech one might say I suck at design like so then you'd want to lean in on design right and You' you'd go to them and you'd say hey like we both agreed on this Tech stuff but like where do you want to go with the design they're probably going to say like please help me and like get creative um same thing goes like you're basically playing the Opposites of what their competency is and so something that just happened to me recently which I like just thought was funny was like working with a very design oriented one and they're like just you know I trust you make the design awesome here's the brand guidelines and then you go and you do the work and you come back and like no like the design should have been like the old one we just wanted these one change like this one change so I think you know even if the design is good like if they're really adamant about it staying structurally the same then like don't do more than the client asks when it comes to like areas that you know they're going to be really touchy on like I think it's just like do exactly what they want and then if there's other areas like you know in my case it was SEO optimization it was small ux fixes it was uh image compression which would technically still be like SEO like that's off page or that's on page where it's like um I actually made a video about this but a lot of people they don't realize but they'll have like a 32 Meg image and they upload it and then it'll try to compress it natively in whatever no code tool you're using but then the quality goes to crap because the algo they're using isn't that great um so that's just something I wanted to cover there the last stuff see already been recording 49 minutes we're going to go for an hour the trouble-free website um yeah we can look at that so I made this website for the community that I'm building and oh my gosh that's awful so we're going to do some quick changes and I guess this should be an example of of vibe coding but this would be something that's really easy so we might as well I you can tell like this just looks awful why is it why is it purple why is it telling me to sign in so let me open this up but this has a bunch of um um it basically is like a a prompt library and so let me find it fre F free AI cool so we're gonna open this up and let's just make it look decent real quick so we're going to do some actual coding and some uiux so let's get this and then we're going to put this on the bottom so that we can keep an eye on it and then we're going to put this on the top going to cover that and we're going to open up the example I really like the stuff that the guys over at midday do so I think it's cursor directory yeah so it's fine um I think this sales page just sucks so I want to on the homepage what would I want to have here I think for now we should just move all this blog content yeah so let's do that let uh let's think of a prompt here so I know that the pages are blogpage blog slash category slug SL page I know blog slash star start nope didn't want to do that I'm just trying to see like what are all the routes that are associated to this so I know yeah let's grab these let's grab those are just components let's grab on the blog itself let's grab the prompts and we're going to come with the plan we'll use o1 and I just want to bring all this stuff that's on here and just it on the homepage because I think it's kind of stupid that we have two separate ones this prompts thing yeah it's gross so okay there anything that I'm missing when we open that that's fine analyze the following files and come up with a step-by-step plan that moves all of the content from the blog to the homepage we don't need any of the content that's on the homepage you can delete all of it and we don't need a Blog route anymore we want all of the content moved to the homepage so we use a planning model which is 01 and let's just fire that off so we're in ask mode ask is uh available for I hit a hard limit are you kidding me let's do Sonic thinking I don't understand how I hit a hard limit but I guess I did so let's let that run I'm going to heat up this coffee okay let's see what I came back with this is a just a two-step prompt where you do plan and then you do EX so real real simple because this is a simple thing so I know the stuff that needs to be done goes back to our AI slop combo earlier but based on the provided files here's a cons comprehensive I have this thing I don't know if it's doing it where I can like zoom in doesn't look like it's doing it for you guys that sucks here's a comprehensive step byep plan to move all the blog content let's move the homepage to display the G cont the blog that seems correct yeah because we have category tabs and then we pass in posts as a prop seems right look categories yeah categories TS content with the key and the value of the prop posts are P passed in where post pass in post okay got it I'm this code base for a second yeah that all seems fine I just forgot if I had put um this in a DV or not I hope I didn't um what would the name of one be it would be beginner's guide yeah okay that's right I'm using content package content content collections got it cool yeah so should all be good that seems fine create a new folder structure sure yep update the blog post okay um upate the page generate this new thing why would we oh because this is an individual page right what it should be so if I go into app I go into blog if I go on the category slog page is this the one I was talking about generate metadata yeah that's what it's talking about to generate a new structure so the difference between that and this is what Rams category SL post log yep that all seems right OG image just removes this yep that's right smart and then what else the alternates for the open graph yep that's all right this is what you should be doing right it's actually reading through it and if you didn't oh that looks right then your so cool category slug yep a wait yep what change here that's just the logic for finding them content collections um yeah literally looks great the structure data should be changed to remove this yep update the categories tab let's look at this kind of already did um under the links for all those so let's yeah link boom this goes yep and then open graph image if you're unfamiliar with what that is it's basically the thing that you show up when you uh like if I sent this to someone as a link it's the image it shows up so that should be removing what oh L just move it okay yep remove old blog posts I don't think that that is correct actually because it's pulling from somewhere else oh that's just the routes du Rags Okay cool so what I do then is all that looks good right and that because all that looks good you know it's going to work or it's going to have a much higher likelihood of working so I'm going to make a AI plan folder and then here I'm going to make plan um MD this is just a reusable thing so great I'm going to turn that into agent mode I'm going say take everything that you just discussed and then put it into step-by-step instructions not omitting any detail into the plan. MD format and markdown in atomic stepbystep instructions for a code assistant to complete well that's going um what I think for this video what I'll do is I'm gonna cut this map thing and I can do this tomorrow so let's this and this because I don't think it makes sense to continue doing coding on this video so maybe tomorrow I can cover some more coding stuff and that'll be super Bas slop cleanup and when I actually do that later this afternoon I'll just uh try to remind myself to take note with what happened but yeah like this even this site right like this is a product of like slop like a lot of it the reason that it looks so bad is because I was like oh Vibe code but it's just you know you take shortcuts and then you wonder why it looks bad so it's apparently wunning plan yeah I think for this video and any of the kind of future videos going forward I'll probably put like like I think I'll do a little bit of coding in all them so I might have like some sort of like code icon maybe like the cursor icon and then other stuff that I talk about is like typically design and then like something to do with money so that like people know um like what what's covered in these videos right so I might have like I'm just trying to think through a thumbnail for this so I think I want to have like my face somewhere so people start to associate it with me right oh here's this guy again you hop into the ether and then maybe like a figma icon I think an AI slop like graphic somehow not like that but it'd be like you know like the title for this video would be like AI slop is killing killing you maybe it's like someone that's like eating a poisonous hot dog or something and they're like oh yeah this is great but then actually like like that's the the AI slop right there and then after they have it maybe like that or whatever I'm not sure what's funny is that didn't work at all isn't that great so I would have been better off just doing this but you know it's Hit or Miss but that's why you need to know how to do it so now I'd just go and I just do this but that's enough for today you know when a little bit over an hour I hope you found this video helpful and if you did please drop a comment below if you made it to the end of this I mean by God you're you're a hero I thank you and and very grateful for your time thanks again and drop a comment about any question that you do have and I'll see you tomorrow