GPT-4o Fully FREE Now & More AI Use Cases

All right, so we just had several AI releases or updates that are what I've been talking about since a while. These tools are getting integrated into apps that you actually want to use, so there's no time to waste. So let's have a look at all the pieces of AI news from this week that you can actually use, starting with ChatGPT. And let me tell you, this is the one that gets me personally the most excited out of everything that I've covered across the past few months. Because now I Everybody has access to ChatGPT4O, the new omnimodal model. No, the system is not available yet. That's coming over the course of the next week. But GPTs are available to everybody. The GPT Store is available to Everybody and all GPTs are now using GPT4.0. But there's more. As outlined in this email that I received from OpenAI, all ChatGPT3 users now have access to the GPT. So it's not soon, it's now. They tweeted that everybody has this now. Plus this is something they weren't clear on so far. All free users actually have access to the Browse feature, which allows you to browse the Internet and pull in external context 3 to 5 links at a time. The Vision feature that allows to upload images as context to your prompts. You can give it a screenshot of a website and then say rebuild me this website and it does it. For example, data analysis, which allows you to manipulate data with the new interactive tables, visualize it and do so much more. And GPTs, which are specialized version of ChatGPT. And if you're interested in this, I actually created a video this week covering multiple prompting techniques for GPT4O. So if you're a beginner, that's a great place to start. But there's also some advanced prompts in there and one of them actually packed into GPT of my own. I'll call it the Custom GPT Builder. If you want more details, it's in that video. But basically you click this one prompt, you tell it what you want your GPT to do, and it works way better than the GPT Builder that they have. Surprisingly, I'll stand behind that and I'll happily do a live stream showing you 1020 use cases and doing before after comparisons that this builder works better. So now everybody's access to this. Even if you're a free user, you have access to the store. You could use GPTs. Now this does leave one plan. Why should I be subscribed to the plus plan. Well, one, the rate limits are way higher. It's super limited. On the free plan, you can only run around 10 messages and then it taps out, which is not a lot. Like that's one or two conversations, 10 messages, right? Secondly, you can't build GPTs, you can only use them. And thirdly, and probably most importantly, the new voice assistant that will come over the next week will be a plus feature. But yeah, I think this is absolutely massive news because with over 100 million active users now, all of them can use the GPTs that you build. All of them can take advantage of the prompts that you have crafted over the course of the last year. Potentially everybody can use this. Now you can show your grandma what you have done with AI, what you're working on, or what you discovered. They don't need a paid subscription anymore to use the best model possible. Absolutely huge for this space. And this is a great opportunity to bring some of your friends in and show them some of the productivity improvements that you've been unlocking week by week by following channels like this. So, yeah, that pretty much covers it. And let's move on to our next topic, which you'll probably won't use. But stick around because I think this discussion is going to be a very interesting one. Now, before I show it to you, I need to establish two things. First things first, this is a glimpse of what the voice assistant from OpenAI promised to deliver. And I think the tutor use case is one of the main ones that really blew people away in there. Right now you have to tutor, but most people think of tutoring when it comes to something like school or university or maybe career progress, public speaking, et cetera. But for anybody who's not aware, I want to bring your attention to the fact that the gaming industry globally already in 2022 is larger than both movies and music combined. Okay. And not just by a bit. In 2022, music generated 26 billion and the movie industry generated 26 billion in total. Okay? Gaming 184 billion. So even though you might not be engaging with games, this is a big deal and a massive trend that is only growing. Okay? So now if you take this massively popular entertainment industry and you combine it with AI tutoring, you get something like this. Tyler1's AI League of Legends coach. Let me briefly summarize what this is about. So League of Legends is the most popular esports on planet Earth. It's one of the deepest and hardest to lear games there is. There's like 150 or 60 characters with unique skills and it's incredibly competitive. I don't say it lightly when I say this game is way more complex and higher octane than something like chess. There's just more variables. They change every millisecond. You need reaction times, you need game knowledge. If you're not familiar with how competitive and toxic this game is, this is like the Nerd Olympics. But every second person is mad and just shouting at each other, especially at the higher levels because they try to climb the ladders. And then what this app is about is a popular streamer called Tyler One that actually reached the highest rank on all five roles. Plus by the way, he got into chess recently and he's a 1900 ELO rated chess player now too. So just an ultra talented gamer launched his own startup now that offers AI coaching in this game. And I think this pulls so many strings together because you're seeing an AI tutor in something that's very popular, very competitive. And the reason I bring this up and cover this is because I see this as the beginning of a massive trend. I think we're going to have tutors for all sorts of things. AI tutors. Okay, so the very first tutor that you can have in here is Tyler1 himself. You can have his voice in your ear and him speaking to you as you play the game, coaching you on how to get better and how to climb this ladder. Sup? Sup? Welcome to Backseat AI, your new in game bud. But already they announced it's going to be other creators that where you can switch the voices. Matter of fact, with AI translation today you can actually translate his voice into Mandarin and have Tyler One coach you in Mandarin if that is your jam. So again, for everybody viewing this video, you probably won't be using this. But it's a fascinating example of what is coming here and what is starting to become available today. People will build this from anything from games like League of Legends, but also chess, all the way to tennis and driving. And you will have these AI tutors that are going to be relatively cheap available for any activity that you might be into. The next tool I want to show you is a game changer for businesses and creators that want to get into short form video. And it's called Munch and they sponsor today's video. Look, if you're an entrepreneur or marketer, I don't need to tell you that staying on top of creating social media content is a struggle. It can feel impossible to create enough content to satisfy the algorithm. Trust me, even After a decade of video production experience and having a team of editors and content managers that are helping me, it's a constant challenge. But this is where Munch really shines. Their platform lets you quickly transform long form content into short form clips, perfect for social media platforms like Instagram, YouTube shorts or TikTok. Now here's the key to actually using it. Well, you need to give it a video that can be chopped up easily. So ideally you want to give it a podcast recording, a speech of yours, a meeting recording, something that is long form with not many edits present. Then Munch can go in and it streamlines the editing process. It uses AI to quickly choose the best clips. It adds subtitles and gives you the final output that is ready to be published on your socials. You can use Munch to create dozens of clips for you to pick from from just one single presentation or podcast. It even includes automatic subtitling and cropping, which are both a must. It effectively reduces the editing time from hours to just minutes. And on top of that, it connects to all top social media sites. You can publish content directly from within their platform. So if you're looking to step up your short form video game, you can click the first link in the description to get started with Munch today. Thanks again for sponsoring this video. And now let's move on to the next piece of AI news you can use, which is a feature that Google actually shipped to your phone. And the feature that I'm talking about is actually a super convenient one that you might want to consider. You can download the free Google Photos app onto your phone. This works for both Android and iOS. And once you're in the app, you do need to give it access to all of your pictures. So sure, there's privacy concerns with that. And the feature that they implemented here is some magic eraser. And just from the name of that, you might already imply what it does. It erases objects in a very intelligent way and it works super well. Let me show this to you on one of these kite surfing pictures that a friend of mine took of me a few weeks ago. Look, it's a fantastic picture. Beautiful Portuguese beach, right? Kite up in the air, wind was pumping. But there's this one kite on the water back here, right? I would like to erase this. Now look, it's as simple as going to edit, then going to tools. And here you have the magic eraser that is brand new in the Google Photos app. Okay, it does a suggestion for you, you could erase that, but actually I just want to Go in here, I'll paint over this. And it's gone. What? It's gone. Simple as that. The edit has been completed and I tried this on multiple pictures. Let me tell you, this works super well. Look, I'll just take this other picture and in here I'll do the same thing. But now we'll have more subjects. Right? So in here let's remove the different people, the different kites. This should work perfectly. How about that one? And then one more here. And now it looks like a lonely beach. Haha. Perfect. So you get the point. These things could be done in Photoshop, but this is just super convenient. It's right there on your phone, it's free, doesn't require a subscription. And I believe that this is a functionality that you do not have in the photos app on the iPhone for example, as of today. But hold up, that's not the only AI related thing that Google is releasing. Matter of fact, there's this new research prototype and brother, what's that? Okay, to be fair, this is not something you can use, but it's also not something you really want to use. Have a look at this. They just released the Chat Director, which is a research prototype that brings 3D avatars and automatic layout transitions to your 2D laptop screen. Have a look at this. What were they thinking? How could they release this? I guess this is a Google Meets extension. Zoom has these. I just saw this recently, but yeah, basically it brings multiple people into a meeting room. This is quite terrible. What happened? I mean, what's going on? Google, Seriously, someone explain. I just wanted to share this. And now let's look at the next tool that you can actually use today, which is canva. And I gotta say the recent update was quite amazing. But it's not all incredible, it's a bit of a mixed bag. Okay, so let me give you the overview of what's new here and which parts of this actually really matter. So look, this is going to be somewhat of a first look. If you want the full overview, they have a separate keynote, a one hour long keynote with all the details of this fleshed out. But basically canva, probably the most popular social media post design tool, had a complete overhaul and integrated a bunch of AI features in the process. And some of these are really significant. So one of them is this, for example, magic design. Okay, and every time they say magic it pretty much means AI powered. So for example, magic designs right here, if I just say Instagram Post for an AI learning YouTube channel, it will automatically create various templates for me. Now, look, this is absolutely a combination of some AI tooling and a lot of templates that they already have, but it mixes really well. This is the nice thing about tools that we look at here, week by week, being integrated back into applications that people are actually using, like Canva, like Photoshop, like Lightroom. And I think this is pretty great, to be honest, because, look, if you don't like what you're looking at, you can just rerun it. And look, these are all AI generated, sort of, right? So the images are AI generated, the text is AI generated, and I'm sure it uses a lot of templates under the hood, just like they had before this. But nevertheless, this is one of the simplest ways to create content right inside of something like ChatGPT. You can't really do this because just with one click, I can go in and edit this text, like so, change out the images, and then I can use some of the other features they have now to work with all of this. What are some of the other features? Well, one of them is actually right here. Resize and Magic Switch. Okay, so with this new tool, I thought this one was particularly good. You can transform this into a document. Okay, so let's say we want to turn this post into a blog post, let's say. Not much information here, but let's just give this a shot. And it's going to use AI to reposition everything, to reformat it to fit the format of a blog post. It's probably using the ChatGPT API in the background to generate some of the text of the blog post, etc. Nevertheless, in just one click, you can go from Instagram post to document. Same goes for PowerPoint presentations. You can just transform these things into each other. And look at that. This is pretty. Need a draft for a blog post right away. Now I can go in and edit it. And here's the next feature they have. If I just select this, they have this Magic write thing here. This is pretty straightforward. You might have seen this in every other app on the world by now. It's just a ChatGPT integration with some presets here. So if I want to shorten this, if I want a custom prompt, I can do that too. It's just all of these features that we know coming together and being integrated into software that people are actually using. Yep, there you go. Works as expected. There's a few more here. Okay, so one that I won't even touch on because you might be familiar already. Adobe Podcast did this a year ago. It's a audio enhancer. Okay. If you have low quality voice recording and you want to edit with Canva, which is a fantastic option for beginners, you can use an AI voice enhancer in there too. No need to round trip to Adobe Podcast and then bring the audio back in. That's fantastic. This one, I was a little confused by it, to be honest, because it feels like this has been in PowerPoint since 15 years. I mean, please correct me if I'm wrong, but hasn't something like this been inside of PowerPoint since forever? Different animations for the different elements, automated. I suppose there might be an AI element to this, as this seems to work pretty well. But then also this is the demo, so I don't know. I suppose time will tell. But let's talk about these two down here, because there's now a video generator and a text generator inside of Canva. Now, the video generator is pretty simple to explain because it simply uses one way under the hood. If you're not familiar, Runway Gen 2 is probably the best video generator. Now it's arguable. All of them are kind of terrible. At this moment, everybody's waiting for Sora. And then there's the Google VO model that has been announced. We don't have access to it. It looks a bit better than Runaway, but way worse than Sora. That's kind of an overview of the AI video space right now. So this uses Runway right in here as they partnered with them. So you can expect those types of results in here, which are decent results. If you watch our videos, you'll see B roll clips made mostly with Runway. When I turn into robots, that's Runway. And then we also use it for various other vis. Now that's pretty clear. So let's talk about the image generator here. And I was curious, to say the least, to see what they did with this, because image generation has become so good at this point. There's multiple models that are virtually indistinguishable from real life. So I was really curious to see what they did here. And if I go here and say a cat with a hat, as I always do, and I always do this one just because I ran this prompt like a billion times. Okay, I might be exaggerating, but you get the point. But I think it's really good to have a go to prompt because then you develop an intuition for what a cat with a hat supposed to look like inside of a good image generator. And let me tell you, this is not the best cat with a hat I have seen. I can't confirm this. I couldn't find exactly what model they're using under the hood. But this looks like a very basic stable diffusion, which is decent. You know, it's good that it's in here. I think it's usable for certain things, but it's definitely not the best. So if you want the best results for this, I think, you know, midjourney still takes the crown. And then there's models like Firefly, Dall? E 3 and Leonardo or Kriya that we are about to look at later on in this video that are still best for this. Nevertheless, nice that they integrated. And I expect this to improve over time because they're actually going down the route that I love to see. They are partnering with various companies to bring their imaging models to themselves. So if you share my opinion and you're not a big fan of the image generator they have, then down here you can see you can integrate Dall? E Imagen, you can integrate an image enhancer and all these other integrations into canva. So let's say I want to use Dall? E, then I can do that. But here's the thing. They don't really state what version of Dall? E this is. Is this Dall? E 2 or 3? And yeah, this definitely doesn't look like Dall? E 3. This is the old one. So, you know, I wouldn't recommend using this one. I hope that eventually it will get some of the top apps like a midjourney integration in here. But this is definitely a great start. Now you have everything integrated with a tool that people are already using to design social media content, create videos, all of this. And you have some of these magical tools in here that are using AI under the hood. There's a few more here like Magic Edit, where you just brush over something roughly. And then it uses AI to detect object and you can prompt on top of that change it. You can Magic Grab, which is similar to the Magic Eraser that we saw with Google Photos, but here you can actually move the object and it just fills in the background for you as you move it to a different spot. And Magic Expand is just a resizing tool that fills in the blanks. So overall, fantastic update. And I should add that you do need a pro plan for most of these tools and I do have one. Actually, I didn't even get it for this video. This video is not sponsored by them at all. I would always disclose that. No exceptions, but. But it's just something we use internally. It's the quickest turnaround you can have with multiple people on the team that use Photoshop, which you might be able to see by the thumbnails we're doing recently. But if you just want quick and easy results, Canva is the way we do it. And now we have all these tools in here, so we'll become even more efficient, which is lovely. If you're enjoying this video, don't forget to leave a like, it really does help the channel. And with that being said, let's move on to the next use case. Okay, moving on. There's two new things that have been released by kreia. If you've been following the show for the past few months, you will know about the different tools that KREIA has, because I do like to cover them. They have a super simple user interface, and the quality of the results you can get with it is actually very high, considering how simple the interface is. So there's two new things here that have come out over the course of the last week. One of them is their new video generator, which allows you to combine multiple images into a video and then you can keyframe them so one thing can morph into another and it works decently well. And then secondly, now they're rolling out a video enhance in their beta version, if you're not aware. As of right now, the state of the art for video enhancement still, I think, is Topaz AI. I use it here and there. It's a really good upscaler for video footage, but it is paid, I believe it costs $300 for a lifetime license, which means that if you're not a video professional, it doesn't make sense for you at all. But now tools like KREIA are about to integrate this into it. But with that being said, let's have a look at this little video feature. I took an image of myself here. Okay? This is generated with stable diffusion stream Booth, which for everybody asking, that is what we do use for the thumbnails on this channel. And it's just me in a spice market. Okay? And then I have the second image of Spider man, which I actually generated in here. Okay, So I just said spider man in a spice market on a sunny day. And then the nice thing here is you have the KREA creator right in here. So you can do something like with smile on his face. And you can see this updating in real time. Yeah, there we go. Now we got pretty close. And now we can generate this video because here on the left, I have one image, and here on the right I have a second image and now if I click Generate Video, it's going to to create the frames in between that and morph the one image into the second. And as you can see, while this loads, this user interface is super intuitive. In the middle you're seeing the generation I had before this. That didn't work so well. Let's see if this one is a little better. This is my second attempt at using this. Mind you, I did use the experimental preset here in the bottom left. So if this doesn't look good again, we're going to try one more before moving on. Because, you know, experimental ones are usually a little trippy, psychedelic, probably not what you want for your typical social media post, but maybe a great idea if you're doing a music video or just playing around. And actually I just noticed this new pop up showed up that I should reload for a new version of the app. And this is when the upscaler here appeared. Look. So we can upscale videos in here now. Okay, so obviously this is not the result we're looking for at all. Kind of just wanted to see these two characters morph into each other. Experimental didn't work. Let's go back to film generate one more time. Okay, this is way closer to what I imagined. Now it's not done rendering yet, so let's give it a few more seconds. Look at that. Now. This is actually pretty good, I gotta say. The only problem is that it doesn't really look like me, right? So that is a big limitation. This original image looks way closer to what I look like in real life. That's why we like using Dreambooth. But here it's a bit different. Nevertheless, it kind of works. And then I could even upscale it if I'm on one of the plans. I believe right now I don't have any plan at all. And you can just try this out yourself, just like I'm right now by logging in with a free account. I think playing around with this makes a lot of sense. If you're into video creation, you can transition things from one to the next with the power of AI really easily with an interface like this. And with that being said, let's move on to the next app. And this will be a super quick one because this is just a quick Suno Update. They released 3.5, and if you're not familiar, Suno is one of the two best music generators. Suno and Yurio are the two most popular ones with the best results. And there was one major limitation here that they crushed with this update you can now generate four minute long songs. I think that's the main point of the update. Plus the tease that V4 is coming soon, which I'm excited for because V2 to V3 was such a massive jump. That V4 sound sounds very promising to me. They also said they have improved song structure and vocal flow and there's a few more updates here, but essentially this four minute length is the main update. So if you're using Suno, you will be happy with this update. Okay, next up, this one also just came out after the cutoff on When I record and when we start editing. We take a lot of time testing and editing these apps, so usually I record on Wednesday and this one came out on Thursday. Midjourney Rooms Very simply explained, if you're a midjourney user, you can now create rooms that you can either make public like the rooms in here, and you can just kind of look at what people are creating. Or even better, if you have a community or if you have a group of friends, you can create your own room and generate in there and see what you guys are creating. So this is really a feature that is aimed at avoiding discord. You can just create a room, set a password for it, and then you and your friends could be generating in a private space without having to open discord at all. Pretty neat. Good to know about. Let's move on. And here's a GitHub repo that I just found out about. Nevertheless, I wanted to really include this in this video because I thought this was super interesting. It was recommended to me by Matt Vid Pro, a fellow AI YouTuber. By the way, if you're looking for longer form exploration of things like Suno or other apps, he does that on his channel. So if you've seen this segment on the soon update, you're like, God, I wish you would talk for 20 minutes, 30 minutes about this one particular tool and not just brush over it. That's what he does on his channel. You can check that out. Anyway, thank you Madvit Pro for showing me this because this looks really promising. I unfortunately didn't have the time to dive into this yet. I will do that. We could even do a separate video on it. I could could see that because this is really a fleshed out version of ChatGPT. This has all the little features that would make the interface more clumsy, but would give power users like me and many of you more functionality. Now you do need the very, very, very basics of development knowledge to make this work. You Need VS code. You need to install a Python package called pygpt. But if you do that, you can run this locally on your computer and use this interface to interact with ChatGPT. As you can see, there's many, many features in here. What particularly caught my interest and what I plan to explore more over this weekend is the ability for other things to plug into it. Okay, so ChatGPT is a pretty close system. You can have your own API endpoints and access them in a GPT, but in here you can actually link to things like Google Search, or you can use various text to speech models just by inserting your API keys in the settings here. Or and this is something I've been asking for in ChatGPT since forever is there's presets and there's a lot of presets that are already in here, but you can also create your own. So look, this is something very similar as to what I share. In the free template that you get on signing up to the newsletter, you have different jobs and different presets for them. The difference is that these are very basic. Look, this quantum physics expert just comes with a very basic system prompt from what I can see here. And here's one that really caught my attention. This is something I also wanted in ChatGPT. It allows you to modify calendars, okay? So you could build an assistant that is set up and customized to manage your calendar for you. And when new mails come in, you can just feed it to it and it can alter your calendar. And this is actually what Matt shared, because what I did is I shared my new GPT with him. That allows you to create custom GPTs in just one project prompt. It's very simple. I'll also link it in the description below. Basically hit the one prompt and tell it what you want the GPT to do and it creates a starting point. And what Matt shared here is that I actually took this, created a calendar GPT and then inside of pygpt here he now is an assistant that helps him manage his calendar. Now again, I just covered this a few minutes ago. I'll have to look deeper into this. Nevertheless, I like to share those things because I feel like everybody can gather inspiration and at least see where the space is going and how the tooling we have is evolving day by day. And that's kind of what the show is about. Right? Okay, moving on. This also came out last Thursday or Friday. I think this is an overhaul of Leonardo AI. We covered it before, but now they have a brand New user interface. A lot of overhauls happening this week, right? So like I didn't even prompt this, it just took the images I had here before. But essentially Leonardo tries to be an all in one studio for AI image generation. Okay, so Canva is focused on being a content creation studio and Leonardo is an image creation studio with all the various features like cutting out backgrounds, using reference images to generate new ones. They always had a variety of models that are really good real time generation. All of these features that we talk here about week by week, they all have in one studio. What I always liked is that they have these fine tuned models, which is the way to go. If you want specific and consistent results, you can pick one of these. This is new though. And they also have this drawing tool just like Create AI. So look, if you upload an image here, there's a few new things in here. The one that really caught my attention is the character reference. And basically here you can upload an image of yourself and create others. I always check this out in new tools like Magic Journey and now Leonardo, because usually they're not similar enough to be used for something like a YouTube thumbnail. What I'll do here is I'll take this older Steve Jobs looking image of myself and see if we can recreate this. And I don't know what's with this animal theme, but I'll just turn mine into a penguin zookeeper and see how this does. Okay, so it should reference my face and it should look just like me. Midjourney is usually really far away because it goes into its training data and pulls out the closest resemblance to me, but it doesn't actually generate images of me. That's why something like Stable Diffusion's Dream Booth is really good because. Because you have multiple reference images and then it recreates you really well, as you can see in our thumbnails. I gotta say this is a bit, it's a bit better than My journey, but it's not perfect. It looks similar. Ish, but yeah, it's not quite there. I like this idea. I should turn myself into a zookeeper more often though. Nevertheless, it's a really simple way to do it right? And I just give it one image. You can do the same with styles or other content that maintains the artistic style or the composition of the image. And you can recreate things like this. And the great thing is that now all the features are even more accessible as this new interface is pretty clean. You can change the aspect ratios, you can change the model to different fine tunes in here, you can create your own fine tunes. And I believe you can now also upscale the images. Yes, exactly right. So the alchemy, upscaler and video generation are all in here now. It's all just easier to access. If you need an all in one image studio, this is one of the good options for sure. And there you go. This is using stable diffusion video under the hood. So the results are. They're okay, I suppose. But you now have this all in one interface, which is great. So that's all I got for today. I hope this was helpful and inspiritive to you. Check out the video about prompting tricks that we uploaded this week. It's a really good one if you haven't seen that yet. And other than that, I'll see you next week.