A (0:01)

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.

B (0:14)

ChatGPT just released its new projects mode. Is it any good? What does it do? Is it just a feature to help you stay more organized or is it something that could change your workflow completely? Completely when using Chat GPT? Well, we're going to answer those questions and a lot more today on Everyday AI. What's going on, y'? All? My name is Jordan Wilson and this is for you. Everyday AI. It is your daily live stream, podcast and free daily newsletter helping us all not just keep up with AI, but how we can actually get ahead and grow our companies and to grow our career. So it starts here. This is where you learn on the daily livestream and podcast. But if you really want to leverage what we're learning, you need to go to your everydayai.com all right? There you can sign up for our free daily newsletter where we recap each and every episode every single day. So when we bring on great guests, we, you know, add in some extra insights there as well as we keep you up to date on every single other thing happening in AI that's going to impact you. So that is how you become the smartest person in AI at your company, at your everyday AI.com, also. While you're there, you can listen to more than 420 episodes. We've talked to hundreds of the smartest people in the world at the biggest companies, small companies. Go learn from them and get ahead. All right, so I'm very excited today to talk about projects, a new mode inside of Chat GPT or maybe a new feature actually. And you might think of it as, ah, it's just something to help you stay organized. Yes, it is, but it's so much more than that. All right, but before we get into that, we're going to start off every day as we do by going over the AI news and hey, livestream audience, got a question on the screen there. Let me know and I'll make sure to address this in today's show. All right, so first, Salesforce is reportedly set to already be unveiling its Agent Force 2.0. So according to reporting from Silicon Angle, Salesforce is set to Launch Agent Force 2.0, an upgraded version of its AI product designed for enterprises, just two months after the original and initial release of Agent Force. So the Agent Force platform enables companies to create and customize generative AI agents that can operate autonomously, working with all of your Salesforce data and helping you in different roles such as sales, service, marketing, commerce, etc. So the CEO of Salesforce, Mark Benioff, refers to this advancement as part of a agent revolution, emphasizing its potential to reshape digital labor for enterprise workers. So again, this is just a report right now, but keep an eye out. We'll probably have that in our newsletter if it does happen today. I do believe it is. We are expecting a big announcement from Salesforce. All right, next piece of AI news. Chat GPT's search just got a little bit better and it's available to everyone. So OpenAI announced as part of its 12 Days of Ship Ms. Or 12 Days of OpenAI that they are rolling out their chat GPT search to all users as long as you are at least logged in. So even with a free account, if you are logged in, you should start seeing this pop up. It's a gradual rollout. So the big thing there is this being rolled out to all users, whereas previously paid subscribers got access in October. But it was kind of two different announcements, both related to search. The other one is Advanced Voice Mode. We'll be getting access to search and a lot of people were talking online like that doesn't matter, no big deal. Yes, that is a huge deal. I said giving Advanced Voice Mode access to the Internet takes it from a cool party trick to an actual helpful AI assistant. So pretty big. Also a lot of new changes in the mobile interface, integrating now with Apple Maps and bringing more of the, you know, like how you would probably want local searches, you know, so if you're searching or if you have Chat GPT set as the default browser or sorry, as the default search engine in your browser, a lot more updates and under the hood improvements that just make it a little more user friendly. All right, last but not least, and it has still been this OpenAI, Chat GPT and Google back and forth, right? So as we got some big announcements from OpenAI, we also got big announcements from Google. So Google has officially introduced VO2, the successor to its previous video generation model. So yeah, we just got Sora from OpenAI, everyone got it. So with VO from Google, well, we got a pretty nice looking demo and a wait list. All right, so with VO2, users can generate videos up to 4K resolution and extend video duration to several minutes, although the maximum duration has not been officially disclosed. So the model has demonstrated state of the art performance and internal testing, surpassing competitors like Sora in video generation quality. And according to Benchmarks that Google did release. So also currently VO2 is only accessible for a very limited number of users creators through the Video FX platform inside of Google Labs. You can go sign up for the wait list. Also alongside VO2, Google also launched Imagine 3, an AI image generation model. So I was actually a trusted tester on that. I don't know if I'm supposed to put out videos as a trusted tester, but yeah, four months ago I reviewed it. Imagine three is really good from Google. I was actually surprised at the time when I got access to it four months ago. I'm like, this is as good or maybe better than Mid Journey. All right, so a lot more information on those things and more in our newsletter. So make sure you go sign up your everyday AI dot com. All right, let's get into it. Let's talk about ChatGPT's new projects feature, how to use it, and also I have three useful hacks that I think you are really going to like. There's even some things that OpenAI put out on its website that aren't actually true or aren't actually up to date. All right, so this is one of those features, y'. All. So sometimes we cover things immediately, sometimes we take a couple of days, right? So this feature has been out for a couple of days. But I really wanted to push its boundaries because although this seems like a small organizational thing, and ultimately this is about providing structure in organization, but there's some best practices and hacks, right, that OpenAI didn't really talk about. That I think is very important if you want to get the most out of this. Because let's be honest, all of the best AI tools, features, modes in the world don't mean anything unless you can actually take advantage of them. And one of the things, and this is something that I think is often overlooked is user interface, user experience, right? Are you able to organize? Are you able to find things? And sometimes that can be actually a pretty big hurdle for people, right? They're like, oh, how do I do this? Where did I put that? Right? And that can actually slow adoption down. So although this projects might seem topical, I actually think it is a kind of ease of use thing that is really going to drive adoption forward. All right, so like we talked about, this was rolled out as part of opens open AIs 12 days of OpenAI or 12 days of ship. Ms. We are almost through now. We have four more announcements, so we'll see what's, what's coming. So, and hey, live stream audience, thank you for, for Tuning in as always. So, you know, get your questions in now. I already see Michael and Brian chatting a little bit about GPTs and using them. All right, So I got some tips on that. Don't worry. Let's start it. Let's start at the beginning, though. What are projects? All right, so essentially they are organizational folders, all right, to group chats, files and instructions. So think of it as a way to help you streamline your workflow and also maintain a little bit better context. And you can do that by sharing files at the project level. All right? And I'm gonna go in toward the end of the show and I'm going to do all this live. And for our podcast audience, I'm going to try to do my best to describe everything. All right? But live stream audience, if you do have questions, get them in now. And we are going to be doing a quick little demo here at the end after I get through the essentials, right? What you need to know, and this is ultimately, it's just built for managing multiple tasks or multiple chats more efficiently, right? This has actually been one of the most common complaints that I've heard. All right, so people don't know this. I mean, maybe some of you do, but we do a free prompt prompt engineering training, right? We call it PPP Prime Prompt Polish. Sorry, y'. All. I like, I know a lot of people are reaching out with the holidays. We probably aren't going to have another free live training for a couple of weeks. Maybe I'll do a video recorded one. I generally don't do that anyways. I've trained live. I lost count. I think we're at about 9,000 professionals on how to use chat GPT. All right? And I've been doing this essentially since Chat GPT came out. All right? So I actually don't know anyone else out there that's probably done as many. I. I mean, unless you work at Open AI, Right? But I've, I've done hundreds of live chat GPT trainings, and one of the most common questions I get, aside from, you know, how does this work? Why, you know, why aren't my outputs good? Right? But it's just, how do I organize my chats? Right? They get wild, right? Are there folders? And the answer was always no, right? And there were actually some, some chrome extensions, right, that, you know, you can modify ChatGPT's behavior, but for the most part, I didn't like those because they slow performance down. There's some great ones, but now we actually have the answer to one of the most common questions, and I'm sure people at OpenAI heard this all the time. A common piece of feedback is, you know, you start a new chat inside Chat gbt, and if you're a power user, such as myself, right, sometimes I'll go through 30, 40, 50 chats a day, especially, you know, when I have the Chat GBT extension set. And that is my default search engine, right? Because then every time I search for something, it actually starts a new chat. So when I have that, you know, I might have hundreds of chats a day. It's actually one of the reasons why I disabled it. So I know we have quite a few people at OpenAI listening. One thing I would love in the chrome extension for ChatGPT have an option to disable it starting a new chat, or maybe just put it all into a project folder by default, right? That's one thing I hate. I'm obviously a power searcher. I'm in front of the computer all day. One thing I didn't like about the Chat GBT Chrome extension is I would have like 70, 80, 90 new chats a day. Because every time you did a search, essentially in a new tab, new window, it created a new chat. All right, so projects at the top level, they are folders, their way to be organized, and you can add files as shared knowledge. All right, we're going to get into that more here in our demo. So here's how projects work. So you can create folders with titles just for clarity. Then you can upload files and add chats for continuity, and then you can set project specific custom instructions. So this part is huge. All right? And yes, y', all, like, I know, is this just like Anthropics projects? You're probably thinking yes and no. I'm going to get to the difference here in a minute between both projects and chat GPTs, custom GPTs as well as Claude's projects versus chat GPT's projects. But the ability finally to essentially have custom instructions that aren't applied unilaterally to all your chats is huge, Right? So people have always asked me, jordan, do you use memory? Jordan, do you use custom instructions? No one knows. Because of its application, in my opinion, is poor. The fact that you, you know, when you start a new chat in Chat gbt, you can't say, okay, I want custom instructions to be applied to everything except this chat. The only way you can do that is by doing a temporary chat, which is not a good solution, right? So I've very, very Rarely used custom instructions or memory for that reason. Because I'm a power user, I'm using CHAT GBT for literally everything and I don't want a single set of custom instructions. So if you're new custom instructions, they can be something simple or something very complex that essentially tells ChatGPT to behave a certain way, maybe to respond in bullet points or to always provide depth in explaining something at an elementary level. Right. But you can see how that might not be helpful if you use ChatGPT for a variety of things in your personal, professional life. If you're working on different projects, having one set of custom instructions that apply to everything, not very good, not very useful, if I'm being honest. Now I will actually be using custom instructions, but doing it at the project level. So now it's actually useful. Now it's actually helpful. Right. I can start organizing all of my chats into, you know, I'm always researching different AI products, different AI releases. Sometimes I'm doing things for shows, right? So I can start creating those project files and then have custom instructions. So finally, this is what I think most of us, when custom instructions were first announced, we're like, okay, this is cool, but there's no way to easily toggle it on or off. Well, now there is. It's a little bit of a workaround technically, but I think this is a big, a big thing here. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for chat GPT training for thousands or just need help building your front end AI strategy, you can partner with us too, just like some of the biggest companies in the world do. Go to your everyday AI.com partner to get in contact with our team or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on GEni. Oh, look at you guys. You guys took the PPP course. Mark said it was very helpful. Marie said she's thinking of taking it again. Yeah, Jackie said many of her students took ppp. Yeah, Shout out Jackie and all the Deaul Peeps. All right, so let's go to who projects are available for. Well, right now it is for chat, GBT Plus, ChatGPT Pro, and ChatGPT Teams users. All right, so free users may be getting access to this soon. This is one of those things, you know. Oh, we'll be, you know, giving out some access, so, you know, we'll see. Also right now, Enterprise in Edu customers are reportedly going to be getting access to this new projects feature in early 2025. So how the heck do you even create a project? Well, I'm going to show you the interface, but for our podcast audience, essentially, on your left hand sidebar now, you kind of have three different columns or three different main sections. So at the top, you have your custom GPTs. All right? Then right below that, you will see the new section, which is your. Which is your projects. And then below that, you will have your normal chats. Okay? Okay, cool. This, man, I love it when this happens. You know, I was working on my laptop upstairs, and now I'm on my. My downstairs computer. Yeah, I do this in the little home office here, not in my actual office. And now my Chat GPT account doesn't want to log in via the browser. That's going to make a demo hard. Right. Even though last night I was logged in all these fun cookie issues. All right, so give me. Give me a second here, y', all, while I try to finagle a few things, but in terms of how you use this where. Where it is. So like I said, you have your GPTs on the top. Now you have your projects in the middle. And then in the bottom, you have all of your chats that are not in projects. All right? And then obviously, you can add files and custom settings within all of your projects. I don't see any limit, right, to the number of projects. I don't see a limit to the number of chats that you can have within a project. The only limit that I'm seeing right now is the number of files in a project. So right now that is set at 20. All right, so if you are wondering, yeah, you can't just upload, you know, thousands of. Of. Of photos or anything like that into projects. You can only have 20 files. All right? I did do some testing with files that were more than a hundred pages, and I was actually surprised a little bit that it did a really good job. I was Surprised. In, in terms of handling the long context, I was doing some, what I like to call needle in the haystack testing, which I always encourage. If you're using this at the company level, whether you are on a teams plan, an enterprise plan, first of all, reach out, hire us, that's what we do, we'll help you. But you need to be running what we call needle in the haystack test, right? So if you're uploading, whether you're using custom GPTs, whether you're using projects and you're uploading files, you need to make sure that it's always working correctly. All right? And with the thing with generative AI, hallucinations are a feature, not a bug, right? And sometimes things get wonky, right? So I, I did some testing in project files with, you know, I think I had about 800 pages worth of files in there and I was doing some testing, you know, hiding essentially small pieces of information maybe unrelated to the files that you're uploading in your project, right? So I had a document that was 120 pages. I hit a little something on page like 110 that was not related, asked it a question, and chatgpt in the projects did a really good job at pulling that information in. All right, so now let's, let's quickly talk about the difference between custom GPTs and projects. So there's kind of some similarities, right? So the similarities are that you can upload files to both, right? And essentially you can add custom instructions. So in, in a way this is, you have two different ways to customize or personalize chat GPT for your liking. All right, the difference is, and so those are the two similarities, the two biggest similarities, where they start to differ is, well, what else do they offer aside from being able to upload files and, and being able to have custom instructions? Well, projects really help with organization, with structure, with ease of use, with workflow. All right, where GPTs excel in advanced customization, right? So with GPTs, even if you're not, you know, great at coding, right, you can go in there and you can use, you can tap into other third party services, you can use, you know, APIs, right? To other software that you or your company might use. There's a GPT store, obviously with millions of GPTs, 99 of them are garbage, but 1% of them are actually really great. So there is a GPT store. And if you are building GPTs, they are much more robust in terms of being able to customize with third party services. So you don't get that with projects. So, you know, there are actually some common use cases where you could do either. Right. If you don't need, you know, a third party API or if you don't need to connect to other services, you might now be faced with a choice, okay, do I build this as a project or do I build this as a GPT? In some cases, either will do. There are some more advanced use cases that I'm going to show you in our live demo that might answer some of those questions. All right, so now let's talk about projects versus projects. All right, so Chat GPT projects versus Claude projects. All right, and you know, I get, I get what people are saying, right? Oh, Chad GPT just copied Claude. Ah, kind of. Right, but then at what point is is quote unquote copying something like a folder, considered copying. I don't know. All right, I'll leave that out out for everyone else. And everyone's playing follow the leader anyways, right? Essentially, when one, you know, big LLM maker rolls something out, generally, if we're talking about Google, Anthropic, OpenAI, Mistral, Meta, etc. You know, Microsoft as well, essentially, they all roll it out. They might call it something different, they might call it the same, right? But there's actually some, some differences between Chat GPT projects versus Claude projects. First, let's tackle the similarities. And it's the same thing as I just mentioned. Well, it helps you better organize starting new chats inside of a project and then giving a project essentially files and instructions. Okay, where they differ. Okay, where they differ. And also you can use canvas mode, chat GPT's canvas mode, which allows kind of inline editing as well as now the ability to execute Python code. You can use that inside of chats in a project as well as in Claude's projects. You can use artifacts. Right? So Artifacts is. They're different, Right. If we're comparing artifacts in Canvas, you know, it's, it's like we think they're the same. They're actually two very different things. Artifacts from Claude is really about rendering all different types of code in canvas. In ChatGPT is all about inline editing. And it's actually extremely useful. People are sleeping on Canvas. All right, but still, you can use both of those features in projects in either platform. Here's the differences. So Claude has a much longer context window at 200k, 200,000 tokens, whereas a big advantage for ChatGPT, which I don't know why Anthropic doesn't allow this. You can't add old chats in Claude into a project, which to me doesn't make a ton of sense, right? What's, what's the use of organizing something if, if you're looking at, I don't know, hundreds or thousands of chats and you can't actually go move them, right? So within Claude, you can only start new chats inside of those projects, whereas in chat gbt, you can do it both ways. You can start a new chat in a project or, or you can go down and find your old chats and move them into a project. All right, so now let's get to some useful hacks, all right? And please keep, keep getting those questions in. I'm scrolling through, just looking through, looking for question marks, right? Because sometimes our comments are, are a little, are a little long. So please, if I get your questions in, feel free to get it in twice and I'll try to answer them all at the end. Rapid fire style. All right, so let's talk about three useful hacks. Number one, OpenAI is technically wrong about this. All right, so on their documentation they say that you can only use the GPT4O model in projects. It's not the case anymore. They quietly kind of rolled out the ability to use the new O1 model. That is OpenAI's most powerful model, the reasoning model. It does this kind of chain of thought under the hood, right? So if you go look on the website, it says as of this morning, please note that chats in projects use GPT4O and the model cannot be changed. So the model cannot be changed still, however, you can actually use 01, so you might not know that. So that's hack hack number one, right? So this is newer information, and I don't Even know if OpenAI actually announced this. Maybe they did on their Twitter, but on their website it is still saying that you can only use GPT4.0. All right, so that's hack one. You can use 01. Hack two, if you need files and 01 add files to chats, not projects. All right, so here's where I'm going to attempt to do some screen sharing and hopefully walk our audience through this as well. So apologies, we're gonna have to start from scratch. I couldn't log into one of my accounts, cookie issue. So I have one of my accounts I don't use as much, but that's fine. So let me tell you what exactly I'm talking about. So we're going to go ahead and we're just going to create a project, okay? So as I was describing earlier, on the left hand side, I have GPTs, then I have projects, then I have chats. So this is an account I don't use as much. All right, so I'm going to click New Project. I'm going to call this Test one. Okay? So again, I clicked the New Project button. I'm naming the project, and the first thing I'm going to see as soon as I click Create Project is the ability to customize. I can upload files and I can type in custom instructions. All right, so now you'll see there is a new kind of interface here, okay? And I have a model selector so I can start a new chat. All right? And it is test one. And I do believe you can kind of use emojis if you want to help it, to help you organize better. You can also have colors, which is great, right? So I can make this Test one a green folder. All right, this is great, right? This sounds so small, but for those people that you know, there's two people out there, people that you know, have a very clean desktop, right? In like zero to, like, three folders or files on there. And then there's people that have 10,000 files. All right? So you might not care about this. For people that have clean desktops, you're gonna love this. All right, so again, you can edit. You can edit the name, you can kind of customize the icon, at least according to color. And then I can start a chat right here from this file. Also, you can. Like I said, you can still use Canvas, you can still use Dolly in here, but this is where you will add files. Okay? At the project level, don't get confused because there is. When you create a new project, all right, it doesn't really say, oh, like this is where you can attach files for the project. I mean, kind of. But you might get confused because the. The first Attach file button you see is to attach a file to that single chat. Okay, so let's say you have 50 chats in this test one project. You need to make sure that you do not upload, thinking that you're going to upload your files in this first area and all your new chats will be able to access that information. You need to do it in this second area where it says add files. All right? Not the. Not the clip icon and add instructions. All right, so I'm going to go ahead. I'm just going to drop. I'm going to drop one thing in there. There we go. I just dropped one. I think this is about 100 page transcript. There we go. Just so we can see how it works. All right? And then I X'd out of it. I should probably not X out. I should probably let it fully load, huh? If I want it to. There we go. And sometimes, y', all, this is a new. This is new. It's buggy. Sometimes your files aren't going to show up right away. Sometimes if you start a new chat, they're not going to show up right away. All you have to do is click refresh on your page and they will show up. All right? And then you have custom instructions. So as an example, I can say, always respond in very short bullet points. All right, so essentially I uploaded a transcript or a series of transcripts, business podcasts. All right? So these are different guests that I've had on the Everyday AI show. And I want you to think right now, what could you use this for? Well, you can use this for a lot of things, right? So I'm going to go ahead and say, you know, what are five tips to use? My delete key stopped working. I forgot about that. Chat GBT for business. Okay? So now if I generally gave this, this same prompt, and I'll probably do this, I'm going to do this actually live and in real time. All right? So if I'm doing this in normal chat gbt, all right, it's. First of all, let's see where it's actually pulling from. It could be pulling from chat GPT's own internal knowledge base, or it could be pulling from the web browse with Bing. All right, so interesting. So here it's just pulling from its knowledge base. All right? So it gave me a lot. It actually gave me a much more succinct answer than I would normally get. All right? But now I'm going to do that same prompt inside this project. So again, I'm having saying always respond in short bullet points. I'm going to add some things. I'm going to say, be. Be funny and direct. Don't waste words. Be Kurt. All right? Sometimes Chad GPT is. Is too verbiose and sits on the fence. I don't want that in my. In my custom instructions. All right, so now we got that. I'm running the same thing now in the projects. I'm saying, what are five tips to use chat GBT for business? So you'll see there's an icon here, and the icon says searching project files. So now I know that chat gbt, instead of using its own internal knowledge base, or instead of using Browse with Bing, it is going to just use the files that I uploaded, as long as those files that I upload answer the query and you'll see right here at the bottom. I hope that they can change or improve this, but at least you see this kind of icon and you can hover over it and it says that it's using my business podcast transcript. All right? And it's no nonsense, right? Whereas before, when I asked it of Chat GPT, it gave me like two or three sentences. It was kind of long. This is short, right? So it says automate workflows, streamline repetitive tasks to boost productivity, data analysis, you Chat GPT for quick insights and, and trend spotting in reports. Right? So this is much more. These are much more specific and I think much more actionable tips than when I actually asked Chat GPT for advice on using Chat GBT for business. All right, so that's number one. But I want to get back to my point of files, right? I promise I didn't get too distracted. All right, so now in this folder, I just added a file. So now if I go back and I hover over here, it says when using custom instructions OR files, only GPT4O is available. All right, and you might be saying, like, I love the OH1 model. I want to be able to add the O1 model and use the O1 model in my project. Okay, so let's go to test two. So what you can do. All right, so I'm creating a test tube project. It's taken a second here. All right, so now I can. You see, I can still use the different modes now and I can just upload the file to this chat. Okay, so I can do the same thing. It's probably going to take me a second and I can say, what are five tips to use Chat GPT for business? All right, and then that's going to start a new chat in test two folder. But because I uploaded the file to the chat and not the project, now when I go Back to the Test 2 project, I can still use the O1 model in different chats. All right, so that is a hack that you can still kind of work around this limitation of still using files in the O1 model. The O1 model is extremely powerful. All right, number three model cloaking. All right, I just gave this a name. All right, so here's what we're going to do. Let's do. I have this image generator, A.I. okay, so I'm going into this custom GPT, all right? And I'M just going to say sup? Okay, so I am using the custom GPT mode and I'm just saying sup. Okay, I'm doing this for a reason. All right, and then the custom GPT response. Okay, so let me show you a couple of things. I'm going to go back to some older chats. All right, Here I was looking up some information on. Let me do one where I'm not using some features. Here we go. Okay, so now I'm going to an older chat. All right? This is an account I don't use a lot. It's experiment experiential marketing copy. So now in the 3dot icon. So again, this is a feature that ChatGPT Projects has that Claude Projects doesn't. I can go to an old chat as long as it's, it's using features that are available in a project. I can right click on old ones and add to a project. All right, so now I'm going to an older chat and I'm adding it to test two. All right, so now here is this GPT chat that I just started. All right? It was auto named sup Chat summary. Okay? When I click the three dots, I cannot move it, right? What if you had some great content and you're like, this chat is amazing. I want to be able to use it inside of a project. I want to be able to organize it. I need to be able to get to it quickly, right? And you're like, oh man, I'm screwed because I use the GPT. Yes and no. All right, so let me show you a little hack. I just call it Model cloaking. All right? I just gave it a name. So I'm going to use this same GPT. However, I'm going to at mention it. So I'm starting a normal chat. All right? I wish I was in my, my normal account here, but that's fine. So I'm starting a normal chat. So I'm in the default GPT4O mode. I'm not in the dedicated GPT mode, but I am at mentioning this same GPT actually. Nope, I need to do this a different way because when I do it that way it's a little weird. So first I'm I'm going into chat GPT GPT4O default mode. I'm just saying hi. So I am starting the chat in 4. 0. Okay. Then I am at mentioning this GPT and just saying sup. Okay, so essentially here's the hack when I'm calling Model Cloaking. If you need to Use a GPT and you also want it to live under a project. Right now, OpenAI in the entire world is telling you, oh, too bad, not possible. I'm telling you, I gave up on sleep to find a way to hack this thing and make it to work for you. Because I actually got a couple of messages on LinkedIn and emails and saying, ah, Jordan, the projects are good, but I use GPTs all the time, so projects aren't for me. Guess what? Here you go. Here's how to do it. Start a chat in GPT4.0, then use your GPT like I'm doing here live on the screen. Then X out of that chat, right? So now in theory, I'm just using the GPT4O again. So send a message, right? So you can use your GPT in 99, but start it with 4o mode, end it by xing out of the at mention. All right, so then essentially what we've done is what I'm just saying, cloaking. Okay, so live stream, audience, here we go. This is the GPT4O only. I can't move it into the project when I quote, unquote, cloak it. And. And I'm going up to this next chat, which is hello, assistant interaction. Guess what, y'? All, Now I can add it to the project. Problem solved. All right, so I have one last tip for y'. All. Yeah, I wanted to deliver. I'm like, man, there's so many good hacks here. There's one more thing. All right, one more thing. I hope that this will be a feature in the future, but this is a thing that people commonly get wrong or a common misconception about projects on Claude and on OpenAI. Right? So if you start a chat, the content of that chat, even if you start it in a project that is not dynamically shared content. Okay, so what that means. Let's say you're. Let's say sometimes I just. I don't know why I use, like, logistics. Let's say you're working at a logistics company, you create a. A project, you upload your company docs, some blog post examples, some SOPs, all these different things. And then you are writing, you know, maybe V2 of your website copy, right? And you're like, oh, this is great in this chat, okay? And then you start a new chat in that folder or in that project. Guess what? Just because they're all under the same folder or project doesn't mean they share information. They only share the information that is uploaded as a file. So a simple hack that I think people are really gonna like, is just manually add finished chats as project files for context. All right, so if you create some great work, some great achievements in a new chat, and you want to be able to share that context amongst other kind of chats in that project, I hope that there's a way that it will do this automatically in the future. But right now, copy and paste that chat, save it as a doc, as a PDF, and then just upload that as a project file. So a simple way to still use what you create inside of a project as knowledge for the others. All right, that's it, y'. All let me go. Cool. People like that tip. All right. Love, love to see it. All right. Yeah. Samuel on YouTube says, I hope the term model cloaking takes off because of Jordan, not because of Terminator reason. All right, let's. Let's go ahead. I think there's a couple of questions here. Let me try to answer them as quickly as possible because, you know, I said this wasn't going to be a long podcast. All right, let's see. Jerry says, what's. Jerry says, what's the difference? Okay, I'm having some, some issues here. Fun times. There we go. Okay. Jerry says, what's the difference between a custom GPT and an assistant GPT? All right, so in assistance, Assistant GPT is more. If you are working on the back end, you're not working in the front end of chat GPT. If you were using the assistance API, essentially you are using the back end. So what we're talking about here is for front end Marie asking, how careful do I have to be with hallucinations? Always be careful. I don't care what if you're using Google Gemini, if you're using chat GPT, if you're using whatever. Right? Always. At the bottom of almost every AI large language model, they say our model can make mistakes. Right. Chat GPT says, ChatGPT can make mistakes. You always need to be aware of hallucinations. Better prompt engineering will cut down on that, you know, kind of quote unquote training. Your chats will cut down on that. Right. The better the input, the better the output. So the more work you put in on the front end, that's what we teach in our prime prompt polish. The higher quality and the fewer hallucinations you will have in your output. Marie asking, I do have some audiobook recordings. Can Chat GPT prep a book for me, example, go through the book, summarize it, give me a character object. Yeah, absolutely. You can do that. And yeah, you could technically do that in a, in a custom GPT or in a project. All right, Samuel, I got an alert yesterday that ChatGPT's memory about me is full. I had to manually delete and confirm a bunch of irrelevant info about me. Any tools like projects for easily managing memories. I'll say this, Samuel, if you are a power user, don't use memories. I don't like it. Anything in there. In theory, if it is important enough, you can put it in custom instructions in a project. I would recommend that. The problem with memories is it, it's kind of your up to chat GPT. Like it does it automatically, right? So if you're talking about something, it's going to automatically say, oh, I'm going to add this to your, your memories, even if you might not want it. Yes, you can go in there and manually delete them. I'm not a big fan of Memories as it's set up right now. Let's see. John, John asking where is the Internet search extension in Projects? Is it not available? That's a great, a great question. So let's see here. All right, so I'm just chatting something here. So yeah, the, the search GPT icon is not available. I can't believe this wasn't on my list, y'. All. All right, so let me just. I'm running a test here live, y'. All. So John, I'm gonna have to get back to you. Uh, unfortunately I can't reply Back to, to YouTube comments. I don't know why, but I'll answer that question in our newsletter today. Mike, Mike, Mike is back. Top guest with the best hair. All right, Mike saying, can we share projects the same way we share chats? So as far as I know, no. Uh, unless you are in, uh, teams. Uh, no, as far as I know. But I will reach out to my friends at OpenAI to see if that is, uh, coming soon. Last question, Nathan asking, once added, can it read all the chat threads or is it just a folder storage? So, Nathan, that was my tip there at the very end. I think you asked that question before my tip. So, yeah, if you have, as an example, 10 different chat threads under the Projects kind of Folder. No, your chat 11 doesn't know the content of chats 1 through 10. It only knows any of the content that you have in the custom instructions or the project files. So that's why I do recommend if you have very important information that you create within the context of a chat, I would go ahead and super hack here, create a Google Doc or a working Microsoft doc and have one file because only 20 files, not a ton, right? You might run out. So what, you could just create one file that's called like chat context and maybe update it every week, right? So it's not like if you have a hundred chats and if you wanted to do that little trick, you're going to run out. So just have one kind of doc that you have for chat context and update it, you know, so you will have to manually delete it and update it to make sure that you can keep using it, if that makes sense. All right, y', all, I hope this one was helpful if it was tell someone about it, right? But again, I think this chat GPT projects seem small, but it's actually a big step to making ChatGPT a little more accessible and a little more easier to use. Sometimes just fixing the user interface or improving the workflow can mean a big difference. So if this episode was helpful, let me know and let your friends know. Let your co workers know, let your sisters, cousins, third grade teacher. No, because we all need to know AI. We need to. It's. It's no longer. I'm an AI enthusiast. Oh, I care. No, understanding and learning AI is an essential skill. So if this was helpful, please if you're listening on the podcast, please subscribe to the show. Leave us a rating. You know, leave. Leave me a comment always in the show notes. You can reach out to me on LinkedIn. Please do love when I hear from our audience as well as email the show. Please click Repost if you're listening on LinkedIn and go to your everydayai.com sign up for the free daily newsletter I promise in today's in today's newsletter, I'm going to answer some of the questions I couldn't get to live. And while you're at our website, go check out everything else you need to more than 420 episodes to learn anything you need to know about AI. Thank you for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks y'. All.