GPT-5 Codex: From Beginner to Expert in 17 minutes

So, I've been using OpenAI's codecs the last few days, and I have figured out a workflow that is probably the best AI vibe coding workflow I've ever seen in my life. The amount of high quality code I'm now producing is astonishing. And this is important for me. I run a startup creator, but has thousands of users and is making 300K a year. I really don't think anyone else is using codecs like this. I've watched every YouTube video about codecs, and no one has shown this workflow before. So, if you stick with me here, you'll not only know how to use OpenAI's codeex with their brand new coding model inside and out, even if you've never used it before, but you'll also learn a master workflow that I guarantee you will 10x your coding output. I'm about to show you things that going to absolutely blow your mind. You've never seen a workflow like this before. Let's lock in, leave a like, and let's get into it. So, here is Open AI's Codeex. And you might be wondering, what the heck am I looking at here? Well, you're looking at a couple things. What makes codec so special is that it is crossplatform and can hand off between the different platforms. So what you see here on the left is I'm just mirroring my iPhone and it has chat GPT open that has codec inside of it. And then on the right here is cursor and I have the codeex IDE plugin. All you

need to do to install the codeex plugin into cursor is just go to chatgbt. com/codex. Go down here and click try in your IDE. And then I'm sure most of you probably using cursor, but you can choose cursor, winds, VS code. Choose that. It will open up the plugin inside that development environment and you'll be set up to where I am now. And then after that, all you need to do is download the chat GPT app. Chat GPT app automatically has codecs as a part of it. It's in the sidebar on the left. And then you have the setup I have, which is the ability to work with codecs on the go, at home, before you go to bed, anywhere you are, McDonald's, Shake Shack, doesn't matter. One other thing you'll need to do to set up before you start using Codeex is you're going to need to connect it to GitHub as well. So, whatever project you're working on, make sure you upload the code to GitHub. That's very simple to do. Just go to github. com, set up a new repository, and then ask cursor to connect the code to GitHub. Once that's done, you're good to go. That's what's going to connect codeex across all the different platforms is that GitHub integration. So codeex is the biggest competitor to claude code is an AI coding agent, but the difference is it allows you to spawn an army of AI agents that do work for you no matter where you are, no matter what device you're on, and have these AI agents code for you even if you're not at your computer. What this allows you to do is basically have an AI army coding for you in the cloud and all you have to do is manage them instead of handhold them like you would in a claude code. Claude Code is an amazing AI coding tool, but you have to handhold it step by step. The difference with codeex and this workflow I'm about to show you here that's going to change everything is I'm going to show you how to create your own AI army that's going to be doing the coding for you. Basically, what we're going to do is we're going to kick off some tasks on our cell phone over here on the left on the iPhone. and we're going to kick off some small tasks we want our AI army to work on. Then I'm going to build out a bigger task inside cursor and then we're going to have it all work simultaneously and put it all together in the end and show you just the output you can get by using codecs. We are going to be working on this AI chatbot. We're going to add some really cool features this AI chatbot I'm building. The ability to compare models and save prompts and optimize prompts and things like that. We're basically going to have our AI army build out all these features at once. So, here's how I like to think about codecs in the different platforms that's on time. For the web interface or the mobile app, I just like to give it small tasks. So, if I'm on the go and I think of a new feature for my app, I'll give codecs on my phone or if I go to chatgbt. com, small little tasks to work on, although you can certainly give it bigger tasks. So, go here. Let's give it a few tasks

and then we'll start working on our own inside a cursor. So, I want to add a new tool up here. prompt library and I want to be able to save any chat message I send to that prompt library. So let's do that. Let's kick this off here. Let's do a new task. Again, I'm just in the chat GBT app inside the codeex option. And we can say I want to add a prompt library tool to the top bar in the tool section. So I'm going to hit send on that. It's going to go right into that GitHub repository we have set up. Boom. task set up. Now the AI, one of the AI agents in the cloud's working on it. But let's keep going. Let's add a few more tasks here before we start building out a bigger task inside cursor in the codeex plugin. I want to add another tool to that toolbar in our app. So I want to add a prompt optimizer tool to the toolbar. In that tool, users should be able to enter a prompt and the AI optimizes the prompt. So, I want to add a tool where users can enter a prompt for the AI and then the AI will actually optimize that prompt for them so they get better prompts out of it. So, we're building our own prompt optimizer inside that tool I showed you. I'm going to hit send and now we are going to have multiple AI agents working on tasks for us. It's like our own employees, right? We basically have two employees working for us here instead of us having to handhold an AI agent like you would in Claude Code. I now have employees working for me. And now I can come over here. I have the codeex plugin inside a cursor. Now I can get to work on my own and start building out my own functionality while my employees are working on other features. And now you can start to see as I do this the power of codeex and why no other tool is able to do this and how you can start getting kind of 10x results out of codeex. So while our agents inside of codeex are building out new tools for us here, I want to add a big new feature. I want to have the ability that when I'm chatting with an AI, I can choose multiple models and I can send one prompt to multiple models to compare their outputs to see which is better. Basically like a prompt racing almost. So, let's do this. Let's go. We're in cursor in the codeex plugin. I'm going to say I want to build this locally inside a better AI, which is the app that codeex over here is working on. So, here's the prompt. I'm going to give the codeex plugin inside cursor. I'm going to say I want to add the ability to choose two models in a chat window. When I select two models, the prompt I hit enter on should go to both of those models in a split screen view. Then I can compare the outputs to the two models side by side. And I'm going to hit enter on that. And now like you would kind of in clawed code, the AI agent's going to start working and writing the code side by side with you. Right? So I'm kind of handholding the agent as we build this together just like you would a cloud code. Same exact workflow here. The difference is now while I work with codeex here to build out this main feature inside of our app, we have other AI employees working for us on the mobile side building out other features. And this is already really cool what I'm showing you here. But think about it from this perspective as the plug-in works. Like obviously I'm sitting here at a computer, but now imagine the scenario where I'm on the go, right? And I think of an idea while I'm in line for McDonald's to pick up a Big Mac. Oh, I could build out this feature. And I give it to my AI employees, right? And I come back and I start working on my own. Or I'm about to go to bed and I think of 10 things I want to build out inside Codeex while I'm sleeping. I can give those 10 tasks to Codeex on my phone real quick before I go to sleep and I wake up and those tasks are all done from the AI agents, right? So, there's so many ways to think about this. There's so many different ways you can leverage codeex being crossplatform and having this independence, right? But the way I really like using it is the way I'm showing you here where I have basically my employee management section, which is how I use my phone here, right? I have codecs up here. I'm managing my employees with all these small little tasks. And then over here on the right hand side is where I'm working with like my senior dev to build out really important functionality. And what I'll show you when all three of these are done is how you then manage the tasks when they're done, how you integrate them, and how this really is actually very easy to integrate all the code from the different AI agents. The main AI agent inside of the codeex plugin is working just like you would a Claude code. All the code changing here as you can see on the right hand side. In a second, it'll be complete. It actually breaks it down to task for itself just like you would see in Claude Code as well. They're very much similar and it's all using the brand new coding model from OpenAI GPT5 codecs which a lot of people love. I'm still undecided on but it definitely is OpenAI's strongest model yet as we can see here while our main model is still working on this big feature we're building out. Uh the task finished the AI agent for the prompt optimizer tool finished. So we're going to check that out in a second. I'll show you how you can integrate all the features that the AI agents built out. That's going to be really easy. By the way, as this is working, make sure to leave a like if you learned anything so far and make sure to subscribe as well. All I do is make awesome videos about

AI. Okay, looks like the task is done. The main feature we're working on. Also, side note, looks like the two tasks we gave our agents on the side to do are done as well. So, our employees are doing a very good job for us. So, what I'll show you here now is okay, we have five employees doing different features for us, building different things out. Won't this get confusing? Won't it be hard to integrate all the features at once? No, no. I'll show you how to do that right now. We'll start off with the main feature we built out and see how that did. Then I'll show you how to integrate the work from our other AI employees. All right, so let's do this. Let's test it out. Uh, select up the two models. Okay, so we can models here. Let's choose GPT40. So we got Llama 3. 2 and GPT40. Let's say, hey, are you there? Let's see if this works. I'm going to hit enter. Yes, I'm here. I'm ready to help. Okay, so they went to both of them. GPT40, Llama 32, both work. All right, so let's give it a more complex task here. Let's say, write me an email announcing a new feature for better AI. Let's see how this goes here. Okay, so it's sending it to both. Uh, okay, it's a little confused because it's opened up four windows here, but that's fine. But look at this. Okay, so we can now compare between two models the outputs for both. So we basically oneshot this feature, right? Where I can send one prompt to multiple models and compare the outputs. So our main AI agent, right, the plug-in for cursor for codecs oneshotted this pretty complex feature, right? That's integrated. That's all good to go. But now let's see how our AI employees did that. We spun up on our phone. Maybe we were in line for In-N-Out. We're getting a burger. I give my employee a couple tasks to do. They're working. I worked on my own thing inside the codeex plugin here. Let's see how our employees did to see how good this workflow really is. So, when I'm in the codeex plugin here, I can hit back and I can actually see here all the tasks I have for my app, including the ones I've done on other platforms. Right. It's also pulling tasks from the Codeex desktop app, jackg. com/codex. I can put in tasks here as well. Right. So this is showing me all the tasks from across all the different platforms I'm using Codeex on. Again, an advantage compared to Claude Code because Claude Code is just the CLI, right? So here we go. Add prompt optimizer tool to the toolbar. Right now it's not there. Right now it's just these four tools right here. So let's do this. Let's pull down the task. If you click on the task, it shows you a summary of what it did. The AI agent gives changed and then it shows you all the code changes it makes. And now what we can do here is we can click apply changes and continue locally. So we hit apply. It pulled down that code. Now let's pull this over here. So this is for the prompt optimizer. I hit refresh. So it's not here at first. I'm going to hit refresh. And there it is. Prompt optimizer right there. So I pulled it down. We go into prompt optimizer. Wow, it built out this entire feature for us. So, this is all brand new. Prompt optimizer transform rough ideas. Let's just see if this works. I'm just add I'm curious if it the AI agent actually built out the end toend functionality here. So, I'm going to say, write me an email announcing a prompt optimizer tool I implemented in my app, better AI. Let's see if it actually optimizes the prompt for us. This would be something special that just end to end built out everything from that one prompt. Wow, look at this. It actually optimized the entire prompt. It added way more to it and it gave suggestions. So that one little task we gave add prompt optimizer tool to toolbar actually built out an entire end to end massive feature. There's a whole bunch of other things. Prompt library. Does that actually work? Wow. Prompt li Okay, so prompt library will be back here. We can go back to the chat. It builds out a pretty complex feature and all we did was quickly go on our app in Codeex and say, "Hey, build this out for me. " So all I needed to do was send these tasks out right on my mobile phone. Then I come back to my desktop. I just click apply changes. It takes all the code changes from the cloud that our agent did in the cloud, applies it locally, and now all I have to do is commit this to GitHub, which is really simple to do. I basically have AI employees working for me whenever I want. Right? If I'm using Claude Code, if I think of something to do, I have to go to my computer, open up my CLI, and give it commands. With codeex, what makes this workflow really, really interesting is it's crossplatform and has as the handoff. So I can put tasks wherever I am on mobile and then continue working on them on desktop. And if you're sitting at your desktop like I am now, you can just pass off tasks for your other AI employees to do inside codeex while you do your own thing and build out your main functionality inside the plug-in here. This is why I think Codeex might be the best AI coding tool right now is because they've built out this really awesome ecosystem of different ways to use codecs, right? You can delegate when you want to. You can delegate the simple tasks. hard tasks. And you can work on the more difficult tasks by yourself if you want to handhold it. But here's even more cool things you can do. So, I showed you how you can have AI employees that code for you. How about AI employees that do marketing for you? I'm gonna go back into codeex on iOS

here. I'm gonna say, "Write me a blog post that announces better AI and all the functionality in it. " Right? So, I'm going here. I'm going to have it be my marketing intern now, and I'm going to say, "Hey, write me a blog post that announces this app we're building. " I'm going to hit enter on that. And now we have marketing interns going and building content for us. Right? If I want to make this a Twitter thread, I can ask for a Twitter thread. So, this isn't just having AI employees that code for you and developer employees. This is having an entire army of AI employees across the board for you. This is how you have a one-man company is by using codecs. It's not just for coding. So, while I'm spinning up AI employees to build new functionality, new features, I can spin up an employee that writes me blog posts. builds me an entire road map. I was playing with this yesterday for Creator Buddy. I had an AI come in here and actually build me a roadmap for different email functionality I'm building out for creator buddy right so you can use these as product managers this is how you 10x your workflow with codeex and this is something only available in codeex at the moment is the ability to build this AI army that is crossplatform across mobile desktop inside cursor anywhere you want and because of handoff I can now go inside cursor, right, and manage that task I just gave on my mobile device. This has exploded my productivity over the last week. I'm mind blown that more people aren't really talking about this workflow at all because this is probably the most powerful workflow I've ever seen when it comes to vibe coding. And boom, just like that, I have a blog post right here ready to go all about the new features and better AI. When people talk about oneperson businesses becoming billion-dollar businesses, this is what they're talking about. So, I'm excited to see how they expand this codeex ecosystem and add on new features and functionality, new ways to use it, new ways to communicate with codecs, new ways to spin up multiple agents, because I truly believe this is the future of how we're going to interact with AI agents crossplatform, no matter where you are. Have the ability for agents to work in the cloud and do whatever you need. If you learned anything at all, make sure to leave a like. Make sure to subscribe and turn on notifications. All I do is create amazing videos about AI. and I'll see you in the next