Massive Week for Google AI & More AI Use Cases

Welcome to another week of AI progression. And we're entering this week from a place of all the big companies competing for the best assistant. We have Google competing with OpenAI. And in last week's episode, I showed you Google's Deep think model, which ChatGPT responded to with their 5.2 model. But then also Google was releasing all of these auxiliary products like Mixport. And this week is no different because the avalanche of releases from these two big players continues. On the the one hand, we have OpenAI releasing their image model to match Google. Amazing. But then also sort of Google opened the floodgates and released like 10 different features and apps all at once. And I'm here to show you all that and more in this week's episode of AI News. You can use the show that, well, keeps you updated in this space race for best AI assistant. Or I suppose the end goal of it would be the race to AGI. But right now we're in the phase of it still just being an assistant, a very capable one, though. So let's look at what changed this week. So the most significant release this week for most people is probably the ChatGPT image generation upgrade. They released ChatGPT image 1.5, and I covered it in a separate video comparing it closely to their main competitor, Gemini's Nano Banana Pro. In that video, I go into several examples and give you my verdict, which essentially is they're probably at the same level. And I just wanted to reflect on some comments off of that video because not everybody agreed that they're at the same level. And it was kind of funny. It's. From what I can tell, it's kind of a 50, 50 split between people thinking that, hey, no, nanobanana is actually better. Their main argument for that was that it performs more like a graphic designer that has taste and actually composites things tastefully. Whereas people arguing for ChatGPT were saying that, okay, it adheres to what you tell it more, the prompt adhesion is higher. So when you tell it to do a certain thing, it does exactly that and doesn't bring a lot of creativity into it. Can be both strength and weakness. But also people pointed out that actually Santa had four fingers in one of these examples. And yeah, that was right. I actually missed that. Look at this. This glove only has four fingers, but it looks more like me, which is a strength of ChatGPT. So if you want to learn more about that story, go check out the video. And hey, by the way, if you're enjoying these stories, we do this every Single week. So subscribe to the channel for more episodes like this and other tutorials. And, yeah, let's look at the next piece of AI news that you can hopefully use. Okay, so next up, we gotta talk about all this Google stuff, because as I mentioned, they opened up the floodgates and they're just releasing feature after feature, feature after app, combining features, doing this, doing that, and some of it is really potent. One of them is actually so potent that I had to create another standalone video. Yes, there's a lot of AI stuff coming out, and I want to give it the time and attention that it deserves. So I created an entire video just focusing on the Gemini Deep Research feature that is only available on their $150 plan that now creates images and graphs and visualizations and all that good stuff. And as a part of the report, if you're not familiar with deep research, quick summary is that it's basically an AI system like ChatGPT, opening up 50 tabs, doing a bunch of Google searches, and then compiling all of that into a Deep Research report for you. And this Gemini version now also includes graphs and visuals and slides and other ways of supporting that research. Now, long story short, this integration is really amazing, works really well, and Gemini has the upper hand here. But you do need the $150 plan to use it if you want the details of the comparison between those two. Well, that's what the separate video exists for. You can go check that out there. I go into a prompt that compares bed sizes for married couples in different countries to divorce rates in the same countries, and the results were kind of amazing. I was surprised by them, too. If you're interested in that, go check out the video. We basically look at that example from different angles. And another one. But yeah, long story short, Gemini really pulls ahead in terms of deep research. That's not all for this week, let me tell you. There's way more. And one of those examples is the them releasing a new model. So you might remember Gemini Free Pro from a few weeks back. This is the thing that really stoked the fire and started this whole competition between Gemini releasing Gemini Free Pro and then OpenAI coming out with their models and all of this madness. But now they came out with a new model. And I know this is a lot. Look, it's basically Gemini's flagship model, their best model, but now in a version that is quick, therefore the word Flash. And long story short, yes, developers can access it, but they also made it available in their assistant that is now extremely quick and rather than me just telling you about it. Let me show you. So let's create a new chat in both Gemini on this side and chatgpt on this side. I'll change the settings of the models here in Gemini to fast. And this is the new Flash model under the hood. Okay, selected that in ChatGPT, I go and switch the model selection to instant. This is their fastest model. And if I say something like, write me a long essay about cats with hats, whatever, I guess that's just my default example. And then I hit enter in ChatGPT and enter in Gemini, we can see this resolve in real time. And as you can see, this is a significant difference. I'm even zoomed in and ChatGPT Gemini is already done. I didn't even manage to scroll down, whereas ChatGPT is still writing the fourth paragraph. It's quick, but it's not as quick as this fast model. So you can see this is still generating and this has been done. I guess you could consider this, I don't know, five times as quick, maybe ten times as quick. I mean, look, I'll just run the same prompt in a new chat and I think this will be done before ChatGPT manages about the same time. Really impressive stuff. And if you need a result quickly, well, I guess Gemini is the way to do that in these consumer applications. Now, obviously it's not just about speed, but as you would expect, this model, which is in their Gemini Free family, is also very capable. It performs reasonably good on all the benchmarks, and most people won't even notice the differences here. If you look at the ALA Marina score, Gemini Free Flash, while it might be quite as high as Gemini Free Pro, it's the still scores above GPT 5.1 and different variations of it. You can see that cluster of the models right here. And just as a quick reminder, these ELO scores are basically users rating it which model they prefer the answers from on like tens of thousands of questions. So it's really a measure for user preference, not the only benchmark, but one to consider. So at the end of the day, this type of speed with this type of performance is impressive no matter how you look at it. And as I told you, that's not the end of it when it comes to Gemini. There's actually one more feature that I want to talk about for a minute because I think this is the most significant thing when it comes to where these apps are heading. And by these apps, I mean ChatGPT, Gemini, Claude, all of them. Google pulled a big move Basically they built in some of the automation tools that you might have already seen, like N8N, make Zapier, all of those. And they built them into their gems. And I know if you're not familiar with all of this terminology, this can get a bit confusing, but basically gems are their specialized like chatbots that you can build. They're the equivalent of a GPT in ChatGPT. And what Gemini did here is in their version of that assistant, you can now do workflows inside. Now here's the thing, I unfortunately don't have access to it myself. So we had team members who are situated in the US who already had access and Noah VPN didn't solve it for me. To test this feature. Look at this interface. It allows you to do multi step processes like hey, take in this video, summarize the video content, take the summary and turn it into instructions for infographic. Then use nanobananapro to actually create the infographic and then turn the infographic into a website that includes the infographic. Doing this with a no code automation tool would have been really hard. Doing this inside of here, it's kind of simple. You can check out the entire process here just happening in real time. And then once you set it up, once, whenever you do this workflow of like turning any video into an infographic website, well you just use your custom chatbot that you set up and it just does it. I think this is such a smart use of these tools and it almost makes me think like, how did I not see, see this as the next phase of these tools? Everybody has been flocking to N8N and its competitors over the last year because it's just so interesting. You can do so much. At the end of the day, I think the value for most people is not there in terms of like what time you invest and what time you get back from those workflows. But once that integrates into Gemini and then obviously ChatGPT will follow and people can kind of just speak and say what kind of thing they want to achieve and ChatGPT will be able to set up this workflow for you. It unlocks that power for the masses, which is just not the case now. I think this is super smart and I really wanted to show you I'll keep my eyes out for this being integrated anywhere else. But yeah, Google is the first mover here, which is kind of a new story that has been consistently happening. Right. They're moving on all of these things, integrating these tools in new ways we haven't seen yet. And I love that they're pushing the competition to up their game. So on that note, there's a few more small stories out of Google this week to wrap up this video we call these quick hits. It's basically where we look at stories that we don't want to review or test, but that are worth your attention. The first one will be a super simple one. It's Google's tool Pomeli, which is like a graphic design tool that is really easy to use. And now there's a new feature where you can like press a button and turn the images and the graphics it creates into videos. So again they're just like combining these features from their ecosystem into apps that are actually useful. The next one from Google that I wanted to highlight is a Google Labs experiment. And I think this is the type of agentic stuff that we were promised by but that never happened, right? Like the idea of a daily morning news brief where AI looks at the Internet, researches stuff for you and customizes it to you. That's like a no brainer, right? Everybody agrees that that would be useful, but Google is actually making it happen with their labs experiment called your day ahead. It's not available yet. You can sign up for early access in the US and Canada, but basically it's an agent that creates personalized news briefings based on what's in your email inbox. And that way you can basically get ahead of your day. Now I think that sounds absolutely wonderful. Other people have been trying to build this. The implementations are like wonky. So I won what this is going to look like. I signed up for this wait list immediately. We're going to cover it once it comes out. On a similar note, another thing from Google and this one is a bit more weird. It's called disco. Honestly, I don't really even get it. I just wanted to show you it's a new type of browser that is also an early access where it creates these generative tabs. So it's like Google Chrome but like it creates all the pages. I don't even know. I think it's a mix between like their Canvas feature and Google Chrome, I guess. And then there's like AI integrated into it. I don't really see where this is heading. They're kind of just experimenting. Maybe someone can figure out a good use case for this. I guess you can create interactive games if you want them. I don't know. This one is also an early access and here they really want you to fill out quite the form to do that. It Actually takes a few minutes because it's so long. But yeah, if you want to create interactive visualizations of solar systems rather than just getting the info in a Google search, then I guess this is their solution to that. And then the last one I promise from Google this week is actually Gemini updating their Google Translate app, but only on Android for now. And what it does now is it reads it out loud. So up until now you could use it in this kind of like weird mode where it like read it out but there were delays. So apparently now it has live feedback and you can just speak to it and it reads it out loud and you can use it with headphones. We didn't get to test it yet, but if you're in the US or India and have an Android phone and you want to translate between English and one of these languages right here, you could try it out now. And concretely it's the Live mode mode that you can access here at the bottom under Live Translate that is supposed to be new. Also looking forward to trying this one more once it arrives on my phone. Okay, so that's it for Google this week, but I have two more quick little stories that I wanted to highlight here. One of them with these AI powered assistants arriving everywhere. Actually on our team we were looking for music and this service called Epidemic Sound. We're not affiliated, but it just worked so well that we wanted to highlight it. It has this AI powered assistant where it can look through the entire library of music just by you expressing what kind of feeling you look for. And it works really well. So it's really amazing to see see all of this tech arrive in all sorts of other applications, including music, search and by the way, also, if you're not familiar, this is not a Newsweek this week, but the Spotify like AI playlist creator works just so well that I'm surprised by it every time. So we're really seeing AI being integrated into everywhere and that's what the show is about. And then lastly, we also see these deals where like Disney is partnering with OpenAI to bring their characters into Sora. So it's just developments on all fronts. And I hope one of these features or apps in this episode resonated with you. I know certainly for me multiple landed. Personally I'm most excited about the gems having the workflows built into them because I can see myself turning a lot of prompts that I use regularly into like little flows and it actually being worth my time rather than where you build the automation. Like I don't know about you, but my experience is that like 90% of the time build it and then you don't use it over time. Feel free to share. We'll feature some of the comments in next week's episode. What you think is the the most exciting story of this week. And with that being said, my name is Igor Pagani and I hope you have a wonderful week.