My New Favorite AI Tool for Brainstorming & More AI Use Cases

Welcome to another week in AI and this week has been very colorful. We have creative tools that you can use for non creative things like brainstorming and ideation. And then there's a bunch of releases from China across the board, including a top tier model that intakes videos as a part of your prompt and you can use it for free. So we have all that and so much more. In this week's episode of AI News, you can use the show that takes all of the AI releases from this week, filters for the ones that matter, and then we get to look at them together. Okay, so here we have an application of which we've seen versions, but every time Google Apps something new, I just get excited because they do some of the most creative interfaces in all of AI and now they're doing something they call a mix board, which is Nano Banana on the back end and a creative interface on the front. Let's just get right into trying this. I'll create a new project. And by the way, if you want to access this, you need to be in the US So yes, I am using a VPN right here. But essentially this gives you this creative mixboard where you can easily combine multiple ideas. So I like this one. How about we make something fun? Show me cats with hats. Okay, now it's going to generate multiple images. So there's a bunch of cats with hats right here. Look at that. It didn't just generate three of them, it made like a whole bunch. Okay, let me upload an image too. How about a taco hat? Yes. Okay, there you go. We have the taco hat. But honestly, this could be really useful for marketing purposes where you just have different branding assets and different products in here. And I really like the way it just created a bunch of them. Hey, this is great. Okay, and then with Shift, I can select multiple. So I could, for example, take this cat and this image and then be like, combine them. And then it's going to use Nano Banana to combine them. Okay, that's not perfect. Let me try different things. What if I just take these four cats right here and I say family picture. Oh. And sure enough, there it is. That actually works super well. Okay. Wow. Very quickly, we're arriving at quite the fun art board here. And while this might be a fun example, seriously, if you put in your logos and just a bunch of different images from the Internet and then you generate a bunch of images in here too, as you saw with a cat with a head, I just got 315 new images. You can start mixing and matching These things. An interface that is not very typical and quite intuitive. At least I find that myself. Create our collage. What would it do with all the images? Oh wow, look at that. So truly this is quite powerful. And mind you, I'm just logged in on a free account here doing all of this. Okay, what happens if I say more like this? Whoa. Three completely different compilations and it's all cats with different hats. Honestly, I kind of like these. I. I might have to print some cat with hat art. I'm thinking this little guy would motivate me to work hard on a regular. No doubt about it. And just so we don't end on cats with hats, I'm gonna go to this plan a party preset and I'm gonna do a cozy tennis themed party in my living room. And ah, honestly this is a fantastic way to brainstorm too. I mean, come on, if you need ideas for something, this might just be one of the best workflows out there. No ideas for decoration, ideas for activities. So I like this. This is not the first angle that I showed you where you can use it to create assets, but you could just use it for inspiration. So now I could maybe take some of these activities like this one, this one, this one, tell it to generate more ideas for a tennis themed party. And now I'm using images to generate new ideas. And I myself as a visual learner would much prefer this from a wall of text that you would get inside of ChatGPT. But if you want text, you can get that too. Like these right here. Tennis trivia challenge. Best match viewing party. Do it yourself Wristband decorating station. And again, all of this is free. Pretty impressive. All right, for the next one, we have a release coming out of China. QEN released six new models and usually you wouldn't spend too much time on this because hey, there's many models out there and China is competing releasing good ones. They're not going to be life changing. You know that by now. But what I was interested in is that if you go to this website chat Quen AI, you have, well all these new models, but also their Max model supports video upload. Interesting. So I'm just going to give this a quick spin to see how that works in practice. I'm just going to pick out some random video from my downloads folder. I'm just going to do this one. It's just me drinking water, doing a few gestures. We'll use it later on, but I'll upload it here and then I'll ask it describe Precisely what you see. Okay, it's going to upload. Meanwhile, we rock out.

Enough. Okay, back to it. So let's just send this and see if it describes it precisely. Wow. So it's generating slowly, but it has a video. And this is the Max model. This is very precise. So it is what I hoped for. Neither ChatGPT nor Claud have this right now. You can find this in Gemini, but this is the first big alternative to it. I like it. Look at that. The scene cuts to medium close up shot of a young man sitting in what appears to be a gaming chair. Dark brown, neatly styled hair, a light beard and a mustache. And is wearing a white long sleeved body button down shirt. The background is softly lit and slightly out of focus. I kind of enjoy this description of an AI of the scene that I see here day by day. And you see it week by week. White paneled doors on either side of him, a lamp emitting warm light, surfboard shaped mirror on the wall. That's not a mirror, that's an actual surfboard. Anyway, you get the point that this actually works. All the gestures are described in here. Arm crossing. So if you need to work with video as context. Yeah, there you go, Quinn. Free Max is your second option next to Gemini. And that's it. If you care to learn about the other models that are good but nothing special, you can check out their blog post. So for the next story, we have something that I was actually excited to try because sometimes we get these video models, but also these animation models where you give it a video and it reenacts it. And these have been getting steadily better. The newest one is from Alibaba and it's called 1.2.2 Animate. Now I did go in here and pay like $1.50 to load 30 credits to try this out. So you cannot do this for free. But what it does in theory is I can upload a Target video, which I will do. It's video that I showed you earlier of me just doing different gestures. Okay. And there's a time limitation on it. So we'll just go from here all the way to here. This should work. Confirm. And now I need to pick a character. Now, during the preparation of this video, we were thinking what kind of character would be good? And I don't know why, but we just landed on Conor McGregor. So I'm going to give it a picture of Conor over here and then it should turn me into Conor. Let's use up 20 credits, which is like a dollar here, to make this Character swap Hap. And hopefully this works. Please try a different image. I'll try this. Okay. It's generating. Should be done in about six minutes, 15 minutes later. So I tried three times with Conor McGregor. It didn't work. So now instead, I'm going to use an image of Rocky because why not actually do the pro mode and generate. So this is Igor Stallone. Let's have a look. Wow. That's unbelievably good. Holy moly. Yeah. So this works. I don't know. This cost me a dollar fifty, but I'd say that's well worth it. Honestly, not sure what else to add. There's other tools like this, but my feeling is this one works even better than what I've seen before. Yeah. Thank you, China, for turning me into Rocky and saving me all that time in the gym. And additionally to this model that allows you to input a video and an image and make the character. They also released a competitor to VO3. It does video generation plus audio generation. And what we figured right away is that, hey, we wonder how well it performs compared to VO3. That is state of the art currently. So we did the most logical thing. We ran the same prompt in both of them. And now let's have a look at what these look like in 1.2.5 and VO3.

This is not just my story, it's ours. If we stay silent now, who will speak later? And in that moment, I knew everything had to change. Not just for me, but for all of us.

So VO3 actually has this, like, robotic tin can. Like, sound is so typical to it whenever you see it on Instagram or whatever. I don't think Juan has that.

It's ours.

It's not perfect, but the voice is better. Let's have a look at another one. Juan 2.5.

It's okay. Don't love the audio.

Yeah, this ambient sound is just a bit better here. None of them are perfect. And then a dragon flying towards medieval kingdom.

Hmm. It's okay. A dragon soared above the castle as the kingdom held its breath. That's epic. And even the video is better. So I can tell you from what I've seen across Twitter and in their blog post, this video model is very good, but it's not quite as good as VO3. But the audio, in some cases, especially when it comes to voices, seems cleaner. It's just good to see more competition in the market. And yeah, if you need a video model that also does audio, 1.2.5 is something you should definitely consider. And as we're talking about video models. There's also another video model, namely cling, releasing their 2.5 turbo API. We ran it through our various tests and as per usual, Kling performs super well. You can see some of the comparisons to the other leading models here on screen right now. Character consistency is good on our little monkey with a boat test prompt here. This is the first time I see the monkey splashing water at the boat, which is very courageous. Water is notoriously hard for all video models. And then we have Mr. Sloth and his cocktail just straight chilling right here in the pool. It's a good model. It's cheaper and it's faster than the ones before. But other than that, not really a breakthrough. So let's see what's next. So for our next story, it's really a combination of two things. It's one news story that went viral this week that relates to AI audio, and two, a new AI audio version of one of the most popular apps out there, Suno. So they released V5. I would call it an upgrade to what we had before. There's more clarity. The vocals sound even better. But comparing it to some of the other models out there, like 11 Labs Music or Yudio, it's becoming really hard to tell what's better. Nevertheless, I want to tell you about the story that happened this week, and we're going to have a quick peek at a song that V5 generated just so you can hear how good this actually sounds. But the story I've been referencing is actually something that hit the mainstream. It's this artist called Ksenia Monet. I suppose it's pronounced. It's basically an American lyricist with an AI alter ego that uses AI to produce her songs. So she writes the lyrics and then she uses AI to vocalize them and create the entire audio. And I got to tell you, I had a quick listen here in preparation for this video and wow, it sounds really good. The lyrics are impressive. You can really tell that these are not AI written lyrics. But have just a quick peek at this. I think most people hearing this on the radio would not know this is AI. We're just going to play a few seconds so this doesn't get demonetized. Really good. The production is incredible. And it used tools similar to Suno V5 here. So with that being said, I guess we're only going to see more and more AI artists. And now let's have a look at this new V5 version. If I just open up my SUNO here, there's various songs that were produced with it. And I will have a quick listen to one of them right here. Let's have a quick listen.

So cozy. You stumbled in. Oh, just look at me. You're making me cry. Oh, I'm grinning, baby.

Honestly, you just can't tell anymore. And I don't think I'm qualified enough to really tell the difference between all these various models at point this, this point. It just sounds good. So if you used to play around with Suno, give it a second look, because the upgrades from V3 over V4 and 4.5 and now 5, they're really audible. It just doesn't sound like AI anymore. And artists realize that. So we're just in a world where AI art is becoming a part of the mainstream and there's no going back. Okay, so to round this video out, I'll be showing you all of this week's quick hits. This is the segment where we look at all the stories that maybe don't deserve multiple minutes of testing and reviewing, but you should still know about them. And we'll start out with a brand new arena that ranks different models. There's many of these, but honestly, I just realized who even runs these platforms like LM arena, they're super popular. But now Scale AI released their own version of this arena and interestingly so GPT5 chat actually ranks as the number one model with Opus 4.1 on the second spot. I always like to get more leaderboards. We follow most of them and let those opinions influence our own usage and recommendations. Next up, we have OpenAI released and it's a new API endpoint GPT5 codecs. It's the model that powers their codec CLI that has become extremely popular over the past weeks. It was the lead story of last week's news. You can use video and now you can directly access this coding focused model directly through the API. Nice. They're really trying to outcompete Anthropic here with their coding focused models because before this their models were the gold standard for for coding. Also, Google is putting AI everywhere, at least they're trying to and seeing how users respond. And that trend continues this week with this brand new AI integration into Google Chrome. So if you update it, you should have this brand new AI button on the top right. For now, this is us only. But essentially you just have this AI button which, by the way, this little sparkle has become the gold standard. Right? That's sort of like the logo of AI across the Internet. Anyway, this little sparkle on the top right opens up various AI options and you can use Gemini from wherever you are. You don't need to open a separate tab. You can interact with the entire world Wide web just like this. And chances are most people watching this video are actually watching it from Google Chrome just because how damn popular this browser is. And now all of it has AI. Notion 3.0 released agents and a whole bunch of integrations. And I think this is actually quite interesting. As you know, I myself am a power user of Notion. I have my entire company knowledge base, all the video production stuff, everything in Notion, even my personal to do lists, and now it all integrates into each other. I think the main thing that they did here is that it allows you to create these different agents with different knowledge documents. And in other apps, I usually export the markdown files from Notion and bring them in here. I don't have to do that. So I'll be exploring this more and more over time. And Notion actually reached out to partner on a video, so we might be doing a deep dive onto everything that this has to offer. I'm genuinely curious to learn all about this myself. And the last quick hit this week is more of a notification, and that is that if you're using anthropic models, up until now they were very clear that they won't train on your data. Well, as of September 28th, that is changing. Now this becomes a setting that you have to disable yourself. So if you don't want them to train on your data, you will have to go into the settings and proactively uncheck the field that says they're allowed to use it. Easy to miss. But now you know, and that is pretty much everything we have for this week. One last note is that there's a lot of rumors about major releases coming up soon. So if you haven't done so yet, subscribe to the channel because we'll be covering all of it. And with that being said, my name is Igor and I hope you have a wonderful week.