Everyday AI Podcast – An AI and ChatGPT Podcast - EP 588: ChatGPT’s Updated Canvas Mode in GPT-5: What’s new and how to make it work for you

Episode Date: August 13, 2025

Have you tried ChatGPT's updated Canvas mode in GPT-5?If you didn't know, most frontier LLMs have a way for you to render and run code in the browser like a virtual computer -- this is what ChatGPT's Canvas mode does. Previously, Google Gemini's Canvas and Anthropic's Artifacts were light-years ahead of ChatGPT's Canvas mode. With GPT-5, though, OpenAI has closed that gap. We'll tell you what's new, how it works, and how to put GPT-5's updated Canvas mode to work for YOU. Newsletter: Sign up for our free daily newsletterMore on this Episode: Episode PageJoin the discussion: Thoughts on this? Join the convo and connect with other AI leaders on LinkedIn.Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineupWebsite: YourEverydayAI.comEmail The Show: [e-mail] with Jordan on LinkedInTopics Covered in This Episode:GPT-5 Canvas Mode Overview & FeaturesHow to Activate Canvas Mode in ChatGPTCanvas Mode: Inline Editing CapabilitiesCanvas Mode: Live Code Writing & RenderingBuilding Interactive Web Apps With CanvasUploading & Visualizing Data in Canvas ModeReal-World Canvas Mode Business Use CasesGPT-5 Coding Advancements in CanvasStep-by-Step Live Canvas Mode DemonstrationsTroubleshooting Canvas Mode Errors & LimitsComparing Canvas Mode to Gemini’s CanvasFive Business Solutions Using GPT-5 CanvasTimestamps:00:00 "Become AI Expert with GPT5"03:46 "Exploring GPT-5's Canvas Mode"08:14 Exploring GPT-5 Canvas Mode13:48 "Exploring Context Engineering in AI"15:11 Live Test: Uploading Large Files17:47 Hosting Interactive Web Game Easily21:42 Creative Career Advancement Tips26:01 Unexpected Features and Mode Toggles28:47 "Podcast Revenue Calculator Need"33:14 "GPT-5 Canvas Mode Enhancements"34:27 Iterative Conversations with GPT538:24 "Client ROI Calculator Tool"42:34 "GPT-5's Impact on Business Efficiency"44:39 "Spread the Word on AI"Keywords:GPT-5, Canvas Mode, ChatGPT Canvas, GPT-5 Canvas Mode, GPT-5 update, OpenAI, ChatGPT updates, interactive editor, inline editing, code generation, code rendering, business dashboard, interactive web app, vibe coding, non-technical coding, prompt engineering, Context engineering, Retrieval Augmented Generation, RAG, interactive learning game, live coding demo, dashboard builder, data visualization, file uploaSend Everyday AI and Jordan a text message. (We can't reply back unless you leave contact info) Start Here ▶️Not sure where to start when it comes to AI? Start with our Start Here Series. You can listen to the first drop -- Episode 691 -- or get free access to our Inner Cricle community and all episodes: StartHereSeries.com Also, here's a link to the entire series on a Spotify playlist. 

This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. There's been a lot of talk since OpenAIs.

GPT5 was announced less than a week ago. A lot of that talk has been on models that were cut off and then brought back and benchmarks and all of these sometimes more technical things, things that we might read on paper. But I think there was one. one low key under the radar, huge advancement in GPT5 that no one's talking about. And that is their updated and vastly improved canvas mode inside of chat GPT. So on today's episode of Everyday AI, we're going to be going over what's new in GPT5's canvas mode, how it works, and most importantly, how to make it work for you.

This is our new weekly segment of putting AI to work for you on Wednesdays. All right. Let's get into it. What's going on? Y'all. My name's Stuart Wilson and welcome to Everyday AI. This is your daily live stream podcast and free daily newsletter, helping everyday business leaders like you and me, not just keep up, but get ahead,

how we can leverage all these new tools and techniques and modes and models to grow our companies and grow our careers. If that's exactly what you're trying to do, this is your new home. Maybe it's your first time listening, your first time watching, your first time watching. Thank you for that. But your cheat code, how to become the smartest person in AI in your department or your company. That's our website, your everyday AI.com. Because there, we have almost 600 episodes you can go watch and listen and learn from the world's leading experts, but also you can go sign up for our free daily newsletter. So we're going to be recapping

today's show with some of the highlights, maybe some things that we didn't get to in chat. GPT's new in updated canvas mode as well as all of the AI news. That's going to be in the newsletter. All right. So like I said, I don't think people are talking about GPD5's new canvas mode. And I'm wondering why. Yes, there's always especially, I would say the GPT5 release, one is one of the most anticipated releases in AI.

history, you could say. I think when you looked at the original release of ChatGBTGBT in November of 2022, let's be honest, there is very few people that even knew that that was coming or knew what generative AI and large language models were. So I do think GPT5 was probably the most anticipated release in AI history, which is when it was released, a lot of the conversation was based around a few things. So it's, you know, according to benchmarks and user preferences, it's the most powerful model in the world. And then people were complaining that Open AI kind of got rid of, you know, old models without warning. And then there was some hiccups on their rollout. And that's been dominating the conversation. But I think probably if you're someone that's just

looking at this new model and you're like, okay, how can I put it to work for me? Canvas mode. Like Open AI actually did a few fairly impressive demos of what their new canvas mode was talking about yet I literally didn't see a single in-depth news article specifically about canvas mode and I was honestly expecting to. So it seems like it was kind of glossed over, but don't worry, we're going to be going over it in today's episode. So today we're going to talk about what's new and improved in GPT's 5's canvas mode. Give five real business use cases of how everyday business leaders can go use it today and show a few live examples of our use cases and things that I'm actually using Canvas mode for, and I hope to use it for more

in the future and just some fun examples. All right. So like I said, we started doing this AI at work on Wednesdays probably about two or three months ago. So this is our weekly look at how we're using everyday AI, how we're using all these new models, all these new modes, all these new features. Primarily on AI at work Wednesdays, we're usually just using either chat, sheep, pt, uh, Google Gemini, um, Anthropics, Claude, or maybe Windows co-pilot. So we're really sticking to kind of the more enterprise large language models, uh, because I feel that's where a big majority of our listeners are spending their time. All right. So there's actually a ton of things that I want to get to, but let's start live. Shall we? Hey, good morning, Dennis. Good morning to

our live stream crew. If you're listening on the podcast, you can always join us live. We go 7.30, uh, live on LinkedIn and other social media, but this thing's unscripted, unedited. And that's why I say when we're starting live, number one, live stream audience, thanks for joining. If you have any questions, comments, get them in now. But we're gonna see because there's a good chance.

Some of these things might not work. I think a lot of times when you look at examples or demos online, they're very polished. They've been run through multiple times. I'm gonna be doing really pushing canvas mode inside GVT5 to its limits live. running some examples that I haven't run before. So I'm guessing at least one or two of these is going to break canvas mode.

And if it doesn't, I would actually be impressed. All right. So we're going to go through some, uh, some demo demons and some bugs, but we're doing it live. So Brian, uh, thanks for joining. Uh, Rodriguez from DC. Appreciate you tuning in. Miguel, uh, join in from, uh, Porto Mont. Uh, so thanks.

Yeah. Uh, and have you guys been using canvas mode in GPT5? Am I the only one? I don't understand it. All right. So let's start live, live stream audience. If you could,

let me know if you're seeing our screen and podcast audience. This is one of those you might want to follow along for, right? So if you're just listening to the podcast side of this, you can go watch the full video because there's going to be a lot happening on the screen. I'm going to do my best to describe it to our podcast audience. But you should probably just go to your everyday AI.com. Click on episodes and then find today's.

episode and then you can watch the video or watch it on our YouTube channel. All right. Yeah, this one might be a little hard to follow on the podcast side, but we're going to try our best. All right. So let's go ahead and share my screen, live stream audience. Sorry, I had to pull up a fresh window of chat GPT. So if you could, live stream audience, let me know once you can see my screen.

I don't want to be talking about things that I'm doing here in canvas mode. And it's not actually popping up. Okay. So a few things, housekeeping things to get out of the way. So GPT5 was released on Thursday. So here we are less than a week later. And there's already been multiple changes to chat GPT. So a lot of them are model names and limits.

Okay. So one thing that's new as of just a couple of hours ago and it does pertain to canvas. And I'm going to tell you why is now there's different model names. Yes. confusing. All right, but now you have auto mode, fast mode, thinking, many, thinking, and pro. Okay, we'll probably go into these modes a little bit more, but auto decides how long GPT5 will think. Fast gives you instant answers from GPT5. Thinking many uses thinking, but quickly in GPT5, thinking normal mode thinks longer for better answers. And then you have,

have pro gpgpt5 pro which is research grade intelligence which you only have on the pro plan but one thing to keep in mind canvas mode does not work on pro i hope they change that in the future because i would really like to see what the combination of the most powerful powerful gpt5 variance with canvas can do uh so for some of these i'm going to be uh going between thinking mode and thinking many okay so if you're following along you're going to start by first choosing, you know, are you going fast? Are you going thinking many? Are you going thinking? And then in Canvas mode, you have to activate it. So sometimes you can ask the model to use Canvas mode and you don't have to click on it. But I always encourage people to click Canvas mode. All right. So what you're going to do is you're going to click the plus button, which it used to be a toggle that was on screen.

So now you have to actually click plus. You have to. You have to scroll down to more and then you have to activate Canvas. So I'm going to show you a couple of things. If you're brand new to Canvas, a couple of things that Canvas can be used for. So it actually has two very distinct purposes. One, you can use Canvas for an interactive inline editor inside of ChatGBT. Let me show you what that looks like. So I'm going to say, you know, write a sales page for every day,

AI podcast by Jordan Wilson promoting a new GPT5 prompting course. Should we update it? All right. So I'm using the thinking mini mode and I have canvas mode enabled. Okay. So the first kind of feature. And for the most part, this isn't new,

but a lot of people don't know how canvas works. And it's actually even this right here, this inline editing version of canvas is extremely underrated. So you have your inline editing kind of variation where it's almost like you are working with chat, GBT in a live document as a thought partner. That's one part of canvas. The other part of canvas is it can run and render different code. Even if you don't know how to code, I don't really know how to code.

I know the basics, but I can't code by hand. So the second big use case of canvas is it can run in render code. So you can have it write code and then it can run it as if, you know, you were setting up an entire website and you had everything technically in order. Okay. So on the first side here, so I did a very simple prompt. I said write sales page, uh, copy for everyday AI by Jordan Wilson, promoting a new GPT5 prompting course.

And you'll see what happens on the right hand side. Um, it created some landing page, all right. Not the best, uh, but this is just for, an example here. Okay. So now it says, you know, by Jordan, you know, Everyday AI presents GPD5 prompting mastery by Jordan host of the Everyday AI podcast. Uh, right. So what I'm going to do right away just to show you an example of how canvas mode works in this one use case. I'm going to highlight this first section here. Okay. And then when I highlight it, there's a, uh, essentially pop up. So it's kind of like you're working in like

Google Docs or Microsoft Word, right? which if you're if you're not using chat g chat gpt in canvas mode you might be like wait you can do this yes so when you highlight something uh you have some formatting options but you ask you also have this thing to say ask chat gbt all right so i'm going to say um you know don't mention jordan's name this is for a landing page uh i'm going to say you know show don't tell or something like that i don't know we'll see what what happens this is more just for an explanation because what you'll see you know, podcast audience. When I do this, it's just going to change and update the section that I highlighted, right? I probably should have switched the mode for this out of thinking, but it actually just updated it in like five seconds. So there you go. And it worked right away. Okay. So now, so not only can you highlight something like that and have chat GPT change something in line and nothing else changes, you can also go down and write something yourself, right? So I can interact in the same document as chat GPT.

So without, you know, let's say only the middle needs to be changed, right? You don't have to, you know, go back and forth and oh my gosh, I only want in the middle couple paragraphs changed or modified, but chat GPT changed the whole thing, right? So now you can just highlight certain sections or you can go in and start chatting yourself. So that's one quick, you know, one half of what Canvas mode is. But I don't think that's where the, the true new power and capability is coming from. All right. So now let's look at the second half of the use cases, which is where we're going to spend

the majority of our time today, which is looking at writing and rendering code, even if you don't know how to code. I think there's so many great business use cases here. All right. So let's go ahead and try this example. All right. Again, we're doing this live. I'm going to keep it on thinking many, all right, because I really want this one to work.

So I'm going to start a prompt. And let me make sure. We should have made sure. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the all-in-one creative AI studio. Powered by Adobe's creative agent, Firefly AI assistant.

assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the assistant. The assistant orchestrates multi-step workflows, drawing on 60-plus pro-grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator Premiere, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching, and creating social variations.

Every step the assistant takes is visible so you can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta. See it today at firefly.adobie.com. All right. I think. Thank you. Thank you, Joe and Angie for saying you can see it. Sorry, Shane. I do a lot of chit chat here.

all right so now let's go ahead and run this in canvas mode so one thing as an example and i did this last week during our a i i work on wednesdays an example of learning the difference and similarities of context engineering which is kind of a new terminology or thought on large language models and RAG, retrieve a long, minute generation. So what I said in this prompt, and you'll see for our live stream audience, it is writing code live. So I said in the area of large language models, I'm trying to learn more about the differences and similarities of context engineering and rag. I said, focus on 20,

2025 info only, then create an informative and interactive learning game in Canvas mode. Make it fully interactive, make it a fully interactive web app that renders in canvas make it interactive informative and some sort of element that makes it either competitive or scoring with progress all right so then you'll see it finished running the code great all right so here's what we're going to do sorry we're going to be juggling around so before we look at that i'm going to try and start another one so i'm going to start another chat and i'm to use the full thinking mode with canvas. All right.

So I'm going to explain what this one is later. But I'm going to show you live because I think this is one of the ones that's going to break it. Because I'm going to upload some pretty big file. So that's another thing that you can do that a lot of people don't know is you can upload different files. So I will let everyone know what these files are. But I want this one to go live because I think it's going to break it. One of the CSV files here is 66 megabytes. It is gigantic.

All right. So I don't even know if it's going to upload. We're going to see it looks like it's going to upload. Okay. Interesting. So all the files uploaded. So I'm essentially uploading a bunch of my podcast and newsletter stats. And I'm asking chat GPT thinking mode to make a dashboard.

All right. So this one, I assume is probably going to take 10 minutes. So it's like a cooking show. I'm putting that one in the oven. And then we're going to go back to our context engineering versus rag game. Okay, perfect. It looks like cake number two is starting to bake, baby. All right.

So now we have our code that is done. So this is our kind of interactive learning environment. That's the difference between context engineering in large language models and rag. So it ran the code. So I have a bunch of code here. And I'm looking at it. And I know the basics of what's happening on the right side, but I actually know nothing else here.

And I think that's important for people to understand when working with canvas. You don't have to know anything about coding, right? You might hear this term vied coding. What's it mean? What's it do? Right. Throw a problem at a large language model and especially GPT5's new canvas mode. And just have it build you something that you can use to learn better,

to be better at your job, et cetera. All right. So now in the upper right hand corner, once the code is done, I can click run a code. And it says run code in canvas. So all right, shoo, it worked. That was a close one. I would have hated for not to work. Okay.

So we'll see here. And you can resize it a little bit here. And then you can also share this code. You can download it and you can copy it because people are always like, okay, well, what would you do with this? All right. If you have someone on the IT side, you can essentially send this code to them. And you really wouldn't need it.

anything else. You could host something like this on your website, on your company's intranet, right? So you can build a lot of different things. If you're uploading files, you know, you have to make sure to upload those on your server and make sure that your, essentially, web environment has any libraries, uh, that canvas mode is, is using, right? But you can also say, here's my website. Here's our tech stack. Can you build this in a way that it will work with us just copying and pasting this file on our server, essentially? So yeah, a lot of people are always like, Well, what can you do? Well, I can share this with anyone and they can use it in the chat,

GBT environment or I can copy this code and host it on my website. All right, but you'll see here, we got a working, looks like it's working, a working interactive game. So the top is this context versus rag interactive learning game. So it says start open the tutorial. All right. This is cool. It's like, it's giving me an order.

So it says to start open the tutorial, then try the interactive modules in any order. Your score in the top right. updates as you complete items. Progress persists in your browser. All right. And then we have a little why it matters, all right, in 2025 and then a suggested path. So it says read the tutorial for five to eight minutes, play the matching game, all right, and then run the rag simulator and take a timed quiz to test retention. I'm obviously not going to go through this all, but we're going to click around and we're going to see if it works. So here in the tutorial, I clicked it. It works. So it's kind of like an interactive tab website with some bright different colors.

So some examples here. Then I can go to the matching game. All right. So let's see what it says. Reranker. All right. So I'm not exactly sure how this game works. All right.

But you can see here, it's interactive and it looks like I have to rank different things. I'm not going to go through and read the five to seven minute tutorial. But it looks like it works in the matching game. And then there's a rag simulator. which is pretty cool. So you can change your knowledge-based size with a little slider. You can change your chunk size, which is tokens. There's a re-ranker and then you can run a simulator.

Very cool. All right. It did it. And then there's a timed quiz. All right. So question one, which best describes context engineering? Designing what the model sees, a specific retrieval algorithm for vector databases or training a new large language model from scratch.

So one of those is rag, one of those is pre-training. One of those is concepts engineering. The first one, oh, I like this. This is a timed quiz too. So I know the answer for this. Luckily, it's the first one. All right. I'm not going to go through all the questions to see which ones are right and wrong,

but there you go. A fully working interactive game right there, right? I didn't have to write a single line of code. It didn't take long at all. All right. So I'm assuming my other one is still going to be working. So we're going to start a new chat here inside GPT5. We're going to go thinking many because I think this one is not going to be too difficult.

All right. So here's what I'm going to do for this next one. So I have a copy of my resume. I took out all my contact information. And I'm going to copy and paste this. My resume hasn't been updated in a long time. And then I'm going to go back into chat, I'm going to paste this prompt in.

All right. I'm going to select Canvas mode. And I'm going to make sure at the top that I tell it to also use Canvas mode. Please create a, I'm going to say using Canvas mode. All right. Here we go. So here's what I put in this prompt. So I said, this is my resume using Canvas mode.

Please create a beautiful interactive website that I could use as a personal portfolio. If you need to, please browse the web to find out more. about me or context of my resume. And that's really it. Right. So maybe you're trying to get a promotion within your company, right? And here's a cool way to do this. Maybe you're a small business owner, a freelancer, right?

And you're just looking at your resume and you're like, man, there has to be a way that I can spruce in this up and do something creative. And there you go. Let's see how long it took. it took six seconds and it looks like it's already done. So I don't think this is going to be super great, but I could be wrong. All right. Let's find out live stream.

Let's see what's in the oven here. All right. So now I clicked run code. All right. It's actually, uh, so I need to zoom out here. I'm a little too zoomed in and things are running together. Uh, all right. This is actually better than I.

thought. Um, okay. Interesting. It's, it's interactive. Some elements of it are interactive. Uh, so you see here, it created almost like a one page website with different tabs. So right now by default, it's on the about tab. Then I can click experience. Uh, and it looks like it should scroll down, but it didn't. There's a speaking thing. It didn't work. And then a contact button, the content scrolls down to the bottom. And it looks like there's somewhere that someone can send an inquiry. So pretty good. But the good thing is, is like when you're working in Canvas mode, just like any large

language model, you don't have to know anything about the code to improve it. So I'm going to say, I'm going to say the buttons. Don't always work aside from the contact button, which works. Now I'm going to say fix that. I'm going to say make it more interactive and cleaner and more visual. And I'm going to say make it dark mode. All right. So I essentially gave it three simple text commands.

And you can do this obviously with anything. I could have done this with the context engineering versus rag simulation. So you only let's see here. So one thing I will say. say in GPT5. And this could be maybe because of some Chrome extensions that I have going on. Sometimes I have to actually manually refresh the page for it to work. All right.

So you'll see here it said failed to generate. All right. So we got our first error. So I'm going to make sure that canvas mode is selected. And I'm going to say at the end, I'm going to say use canvas mode to update this. All right. So here we go. Take two.

Again, I should have first. disabled some of the, some of the Chrome extensions that I had going because I know that sometimes those can cause some issues rendering the code. I actually know for a fact because I've had to troubleshoot this before. All right. We're going to give this one a minute.

Live stream audience. If you have any questions, I would love to hear him. So Daniel said, what happened to Canvas mode in the chat, GPT app? I don't see the option for it anymore. Yeah, I saw that as well. I'm guessing that they're going to fix this. All right.

Michelle says love the extra tools. It offers plus the coding. Yeah, same. All right. So here we go. Let's see if it updated our portfolio website. We're going to run the code. All right.

What do you think, live stream audience? Second version, a little better, right? A little better. I like this. So it gave us dark mode. That works. Everything else looks similar. does look a little cleaner, at least for my taste.

Let's see if the, kind of the buttons work. So, all right, the about button, kind of scrolled it, the experience button. Now it's scrolled it in. So if you've ever used a one page website and there's little buttons and then you click the button and it scrolls you too, that's what's happening here. So it essentially turned my resume into an interactive website. So now those buttons are working. That's great.

It looks like there's a toggle. for light mode versus dark mode, which didn't work. But again, if I really wanted to, uh, to do this, I could. Um, there's also pretty cool. I didn't know this. There's also some little pop-up modals. So I can click speaking. And then there's a little, uh, there's a little, uh, kind of pop up that comes up

about different talks that I gave. Uh, and then you can, uh, email or request a booking. So pretty cool. I didn't even ask it for that, uh, that functionality, but it went ahead and built which I think is pretty cool. All right. So we have a couple versions. I don't know if our other one is going to fully work.

I'm going to refresh it here. So our live stream audience can see it going. We might have to check in on it at the very end. But I do have one other use case that I've kind of like pre-built. And this is something that I've been using. Right. So I don't know if that one is going to work like I. said, I'm using the highest level of thinking.

And I threw a ton of data at it. So one thing in Canvas mode, you can see right away, it put the content of the spreadsheets in the actual kind of view, right? So I can click to expand this table. So these are kind of our Buzz Sprout stats. All the, my gosh, these are all the different cities, apparently that our podcast has downloads in more than 20. 7,000 cities.

I didn't even know there's that many cities in the world. All right. So you can see here it instantly converted all of my spreadsheets and all of my data into interactive tables inside of chat, GPT. All right. This would be a bummer. So it thought for 12 minutes. And it looks like it failed to generate.

All right. I'm going to say, please regenerate. So again, I thought, actually, it looks like it. It's still working. So we'll have to check back in. We'll have to check back in on this one. All right. I'm going to say, you know, please update.

All right. So we broke one. We broke one. I knew that we would probably break one doing that live, you know, less than a week after chat. GBT came out also or after GPD five came out. Also one thing to keep in mind, uh, GPD five just a few hours ago did update to this new, uh, kind of model selector model pickers.

I'm guessing a lot of people are going in to test out the different modes. So probably not the best time to really push chat chbt to the limits, but hey, I broke it once. All right. So one other use case that I actually use this all the time, right? And if you listen to our podcast, I said in 2025, non-technical people are going to be essentially vibe coding different little apps for themselves.

So as an example, this is an app that I use all the time. All right. I don't know why there's no great podcast revenue calculators online, right? I found one like three months ago and I didn't save it. And now I can't find it. So, you know, for me, right, we have a lot of companies who like to advertise with a podcast, right? We recently had Square on as an advertiser. We've obviously had Google on.

We've had Microsoft on, right? And a lot of times I'm trying to figure out what is the right price. Right. So it's hard. You know, our stats are always changing. Right. So I'm always like, wait, I'm not good at this math. There's a formula with a CPM.

And I'm always trying to find the right calculator online. And I just can't find it because it's a very niche use case. Right. And a lot of these calculators online are actually for the media buyers, right? Because it's like put in your budget. It's not for the actual, you know, people with the podcast who are trying to find the right price for ads. So I just literally built, built myself a simple podcast revenue calculator where you can put in the CPM, right? So we might usually do 50.

All right. And then I can sort it by per episode, daily, weekly, et cetera. But then I can go, you know, I can say, let's see, daily. I don't know. We probably have, I have to look, probably about 11,000 downloads a date. We have three ad spots. And if we fill them all, 100%. So there you go. So I can go in very easily say, hey, according to, you know, fair market prices out there. Here's about what I should be charging, what we should be charging, advertisers. And then we can also explain it. Right. We didn't pick up a random number. Nope. Here's, you know, industry averages for CPMs for the type of audience that we have, etc. So this is literally a way that I'm already using GPT5, their new canvas mode. And I'm probably going to make some updates to it as well. But very simple way that you can use it.

All right. So let me, let me double check. Yeah, I think we broke the one. I'll check in on it at the very end. But I do want to, before we wrap, if you do have any questions, get them in now. Yeah, a lot of people. So Robert said still impressive results with decent response times. Yeah, I agree.

I'm loving canvas mode in GPT5. All right. So let's go back here. And I want to talk a little bit of the details, right? We started out live, you know, took about 15, 20 minutes to go over some actual use cases that we're using here at Everyday A. I, but I want to talk now specifically some of the details in the coding advancements. Because I would say, if you look at GPT4O and GPD5, some of the biggest jumps are in coding, right? So you pair that with this new and improved canvas editor.

And I think this is vibe coding 101, right? If you don't want to go, you learn cursor or pay for tools like lovable, et cetera, for the 700 million weekly active users, you have a vibe coding basic platform, right? I wouldn't recommend you try to build an app, right? An MVP for a SaaS or something like that inside Canvas mode, but you can at least get something started. But let's talk about some of the coding advancements in GPT5. So their latest model demonstrates significant improvements in coding capabilities,

surpassing previous iterations and competing models on various benchmarks. One of the big ones was Swee, Bench verified, which is essentially a software engineering benchmark. And it achieved a huge increase, about a 75% accuracy, whereas GPT4 was 30%. Right. So if you were previously using Canvas mode with GPT40 and, you know, trying to run and render some code and you're like, this isn't that great. Well, yeah, it should be now more than twice as good, which is again, why I'm like scratching

my head and I'm like, why aren't more people talking about Canvas mode inside of GPT-5? You know, if you've been listening to the show, I've been bonkers for Gemini's version of this, Gemini's Canvas mode since it came out, and they even had a low-key update where you can use Gemini's AI features in the kind of AI apps that you built. I would love for GPT-5 Canvas to offer that, but right now they don't. So you have enhanced code generation because GPT5 demonstrates stronger capabilities and generating functional and syn synthetically correct code, syntetically correct code snippets in response to user queries, and there's improved contextual understanding. So as the base model improves and understanding your queries, combining that with better coding under the hood and a improved canvas interface, right? It's really just helping you better have an iterative conversation to build something that you can use.

So the example that OpenAI gave in their GPT5 live stream is this engineer at OpenAI essentially created a gamified version to learn French, right? And there was kind of like flashcards and quizzes and also a little game, right? So you can really direct it. You can have an open ended prompt and just say, hey, here's a bunch of data, kind of like the one that I broke GPT5 with. Or you can say, hey, here's exactly what I want, which I like. So normally when I'm building like a business dashboard and interactive kind of strategy guide based on my data, I'll usually just let GPT5 Canvas do what it wants to at first.

and I'll see what it creates. And then I'll either iterate that or I'll start a brand new canvas based on what it built me and then ask it for specifics. Right. So again, seeing examples and what people have built with it and iterating on the original is going to give you the best results. Speaking of use cases, all right. So I'm going to give you five different use cases for the new GPT5 canvas mode. So one is a contract comparison tool, right? So again, what are these things that you think?

I gave you an example of some, some little things that take up my time, right? When I'm trying to learn something in a way that is actually ingrained, right? I love using notebook LM, their new video overviews, but I'm using Canvas mode, right? I'm a very, you know, competitive person. I like to gamify things. So before I do a podcast on something or if I'm interviewing an expert, I might go into GBT5 Canvas now and create an interactive learning guide. So I can really, when I'm interviewing someone, I have a better understanding.

So think of what it is that you struggle with on a day-to-day basis or with the software that you have. It's something clunky, right? So you could create a contract comparison tool. This is use case number one. So again, you have to make sure that you have the permission to upload any documents. I'm going to put that out before I go of these five different use cases. But what you can do is you can upload two different vendor contracts or service agreements and then have a side-by-side dashboard that looks at the different liability clauses, payment terms, potential red flags, et cetera.

And this can just save you hours of manual reviews, right? So that's use case number one. Use case number two, build a meeting summary builder, right? There's all these apps out there, right? Some of them have okay free plans. Some of them have, you know, paid plans. But on the desktop app on Mac, chat GPT actually has a like a meeting function, right, where you can kind of record meetings.

Obviously, you should be getting consent of the people on those meetings, right? But then you can just literally use canvas mode and have a meeting summary app. So you can upload those transcripts that chat GPT can record. And you can build an interactive action item tracker with assigned. owners, deadlines, priority levels, and automated follow-up reminders, right? Pretty cool. Yeah, you might already have systems in place for this, but if you don't and if you, maybe you're a small team or your team is heavy chat GPT users, there's a way to do it right

away. So instead of just a follow-up email that you send to everyone, remember, again, you can share these creations in Canvas mode, right? So you can create instead of a follow-up email, that's kind of boring and dry and, you know, not fun to work with. You can create an interactive visual dashboard, right? If you have one big important meeting every single week, I think that's a great way to look at it.

All right. Number three, use case, a client ROI calculator. So you can upload your service pricing and your client's current metrics. And then you can have a GPT5 canvas build an interactive calculator where prospects input their number to see, uh, input their numbers, uh, to see projected savings, revenue games, uh, and a payback period if they use one of your certain services, right? That's another thing. You have all these, y'all, like in the early 2000s,

there's all these essentially online calculators that became like multi-million dollar businesses over the course of multiple decades, just because there's niche use cases that people always want calculators, right? Like sometimes it's like, okay, I need to upload this document and count, you know, the word frequency or something like that or, you know, a comparison tool, right? Here's a great one. You can build this for your company, put the different, you know, whether it's hourly by the project, you can put in all your prices and then you can create a front end calculator

for your clients or potential clients. Because as an example, if you know that one of your USPs as a business is there's a certain product or service that you provide a better value versus your clients, you can literally show people that with an ROI calculator that you could host on your website. or you could send it to clients. All right. Another one, proposal generator. Literally, you can just build software, right?

There used to be software. I remember paying, I don't know, like a couple hundred dollars a month for a certain piece of software that was just for proposals, right? Obviously, you're not going to be able to have a full-fledged piece of software with one prompt. But if you want something that's better than just sending, you know,

a bullet point email or a PDF Word document that's a proposal, you can create something interactive. You can create something highly designed and with a great UI, UX, right? So you could upload your service catalog, past winning proposals, client requirements, and GP5 can auto generate and build you proposals with customized pricing, timelines, deliverables, and terms based on your client's needs. All right.

So, yeah, in that, you can cut down proposal time from maybe many hours, multiple days of back and forth, right? You can even have a form where people fill that out, pass that information to GPT5 Canvas or, you know, use it within GPTs. That's another one, right? GPTs, custom GPTs, you can have canvas mode enabled in them, right? So that's another kind of huge hack. I didn't even probably should have started with that.

All right. And then use case number five, a budget variance dashboard. So you can upload monthly or quarterly budget versus actual spending. and have GP5 in Canvas mode build a visual dashboard, showing variances by department, category, and timeline with drill down expectations and forecasting. So in this case, you can spot budget issues early, make data-driven decisions, and improve financial planning accuracy. All right.

So those are our five use cases. You know, whoever said someone on LinkedIn, I can't see who said that. I said, I so appreciate your bravery in doing this live, Jordan. Yeah, doing live demos is not easy. Right. But I mean, this is one of the reasons why I do everyday AI the way I do it. I think so much out there, it's pre-published. It's scripted.

It's edited. Right. But that's not how generative AI works sometimes. Sometimes you ask it to do something and you run into some hiccups, right? Um, which is what we ran into when I tried to get it to, uh, do the, uh, my business dashboard with all my data. But you know what? I'm going to rerun that later.

And I'm not going to use the full thinking mode. I'm going to use thinking mini. So, all right, if you want to see how that one turned out, I'm going to share it in the newsletter. And you can go look at that interactive dashboard that I built. All right. So that's a wrap, y'all. So to recap, because, like I said, I think a lot of the recent attention and buzz

around GBT5 is on benchmarks, is on all this drama of they got rid of the old models and now they're back, right? People wanted the, you know, these models that just agreed with them like GPT40, even though GPT5 is better, right? And then there was all these issues that Open AI had rolling it out. And hidden in all of that is, I think,

the best feature for everyday business professionals And it is turning your time consuming manual, at times monotonous decision making projects and tasks into something a little more interactive into things that are going to save you time and ultimately save your business or your department money. Right. So obviously there were huge games in GPT 50 when it came to coding, when it came to to software development. And you might look at that and say, okay, well, that's great for the people who build

software. And I say no. Right. You saw an example. I'm building little pieces of software almost every single day. Uh, right? And I've really enjoyed, uh, although I love using Google Gemmonize canvas. Uh, I've really enjoyed these updates in GPT5 canvas.

I think, uh, this gives a new set of. tools to hundreds of millions of chat chagpt users, but so few people know about it. All right. So now you know how it works. And you can put the updated canvas mode in GPT5 to work for you this Wednesday and every day. All right. I hope this was helpful, y'all.

If so, don't be a jerk. Let someone know about this. All right. If you're listening on the live stream, if you found this helpful, people are always like, oh, Jordan, every day. has been my cheat code. And I'm like, well, tell someone about it, that, right? I think there's so much garbage out there in the world of AI, smokes and mirrors, marketing,

people just try to sell you crap. I want to make ongoing, updated, factual AI education for the masses free. And it only works. It only works if you tell someone about it. So if you're listening live, please repost this, tell someone, send this out to your company. you know, when we send out today's newsletter for this to someone that needs to hear it. If you're listening on the podcast, appreciate your support. I know this was a more visual episode.

So like I said, if you want to see the video version, you can go check out our YouTube channel or go to our website at your Everyday AI.com. While you're there, make sure you sign up for the free daily newsletter. We're going to be recapping the highlights of today's show. If we get that dashboard to finish, I'll share that out there as well as you can see everything we got going on under the hood. So thank you for tuning in today. Hope to see back tomorrow.

and every day for more everyday AI. Thanks y'all. Meet Firefly AI Assistant. Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome while the assistant accelerates execution.

Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your EverydayAI.com and sign up to our daily newsletter so you don't get left behind.

Go break some barriers and we'll see you next time.