ChatGPT Just Changed Forever...

It's a big week for OpenAI. Not just that they released a brand new flagship model, GPT-5.5, but also they have a brand new image model. Create social media graphics from this thumbnail in 3 sizes, Instagram square, Twitter banner, Pinterest pin. And look at that, Instagram square, Twitter banner, Pinterest pin, all from one prompt. So there's really something for everybody here. Plus a few more interesting AI releases that you should keep an eye on in this week's episode of AI News You Can Use. A format that I recently reworked to be simpler and to focus on the things that I actually care about rather than rounding up all the AI news stories that you could put to work today as a non-technical user. Let's start by looking at GPT-5.5, because this is the big one. And let me be upfront, people, including me, are still figuring out if this is going to make a difference in their workflows and the ones that they already have, because this model is not optimized for chat. Igor, a new ChatGPT model that is not optimized for chat. What does that mean? It means that AI is shifting. My first impression of GPT-5.5 is that it is, it is different. As I talked about in an upload earlier this week, showing the 3 different levels of AI, which I'll be referring very often from here on out: AI for answers, AI as a work partner, and then AI working for you. That top level is the agentic level, where AI is doing things for you. And this model has been developed for exactly that. It's made for long-running tasks that are complex. I mean, I cite, it's designed for complex work like writing code, researching online, analyzing information, creating documents and spreadsheets, and moving across tools. Now sure, this happens in ChatGPT too, but primarily these are agentic tasks. Agentic coding is its headline capability. And they keep reiterating this ability to take super messy inputs and turn them into structured outputs that actually make sense. Meaning you can just bring a bunch of notes and say, hey, turn this into a presentation, and it will do it for you. Now, I'll show you some examples in a second here. But across the internet, people aren't sure. Just like with Opus 4.7, the reviews are mixed. And I think the reason for that is we've sort of plateaued in terms of user-friendliness in chat applications. And when it comes to writing code, the bar is so high already with such good models out there that really, a lot of these new models just feel like a different flavor of really good. But rather than just reciting what people say across the internet, let me actually show you some examples of this in action. And as always, I'll stay as transparent as possible. This was the very first prompt we ran. It's something that worked in Opus 4.7 when that came out last week immediately. And Opus 4.7 nailed it. GPT-5.5 thought for 3 minutes and 12 seconds. And it created this YouTube repurposing application, which actually ended up not working on its first shot, whereas And Opposite did. Oh, what a loser. How do you access it in chat? Well, that's actually quite simple. You just go here to your model selector at the top. And by default, you'll probably be on Instant. And this model now hides behind Thinking and Pro on paid plans, plus Pro Business and Enterprise. And you can see this clearly if you go to this configure option. And you'll see that Instant runs on the older model 5.3, Quick Replies, and the new Thinking and Pro models if you pay for the $200 Pro plan. Are available in here. So if I switch to thinking, I can actually pick the thinking effort. For the testing examples, we defaulted to extended because also Opus is not like a heavy thinking model that goes ahead and thinks for 5 minutes. So we tried to pick similar settings here. Now, here's an interesting one. I always run this prompt, right? Create an SVG of the Death Star in the sky above Los Angeles, just to show its ability to, I don't know, visually represent something. And this is the first model that is actually, hmm, I don't know, maybe too smart for its own good. Look, every time I try this, it starts thinking through this process. And it's like, hmm, do we really want to make an SVG SVG. Maybe I should make an SVG manually. Or just make it, come on. In this case, it gave me a path to file I can't even access. In a previous testing, it sometimes switched to the image tool and it generated a PNG of what I asked for. And then when re-prompted, it got this data path again. So not perfect. It's sort of like overthinking it. Again, it's the point that I started with. I don't think this model is made for chat. And that's fine. I will be testing this with my Agentic setup. I want to see if this performs with my OpenCloth. But in chat, I don't know, I don't think I've seen a release where out of the few basic test cases that we run, as I just show you, multiple don't actually work. Now I want to be clear, this is not me saying that this is a bad model. For example, we run another test prompt where it builds this space shooter game. And this is by far the best space shooter I've ever seen out of an LLM. Look at the complexity of this thing. This actually looks great. 3D space, look at all that. So clearly it's doing something right. But what I am saying is that I don't think this is a model made for chat. And my conclusion on this is, hmm, I think OpenAI is getting ahead of its current state of ChatGPT. And what I mean by that is that in the next week or two, or I don't know, maybe three or four, but definitely shortly, I think we'll see an overhaul of ChatGPT and it will turn into something that is a bit more agentic because clearly that's where the entire space is moving. And they are the prime consumer application. And maybe they're just putting this model in there now as a first step. If you're making stuff, it's great, you're gonna love it. For everyday tasks, it's similar. But for some of these specialty tasks, it just seems like ChatGPT is a bit too limiting for what it actually wants to do. Okay, so that's GPT-5.5. But if you're using ChatGPT, there's another big release that happened this week. And this one is actually exciting. So, you know, they had an image generator. Now there's a new image generator, it's called GPT Image 2. And it's so fun. First and foremost, what I tested right away is How good is it at retaining people's faces? This was a big thing inside of Nanabanana and why it was so popular with people. You can upload an image of yourself, like me right here, and say something like, create professional headshots of me. And then look at these. All of them actually maintained my face. They didn't try to regenerate another person who kind of looks like me. This is actually me. Big win. And this is accessible to all ChatGPT users, including free accounts. Really great stuff. Now, a few more things to be noted here. First of all, I believe this is the first imaging model from one of the big providers that allows you to turn on thinking mode. So in practice, if you enable image creation like so, you could also go up here and say thinking, and all of a sudden, it will think multiple rounds before it goes ahead and creates the image, which can be really good with, well, something like animal infographics right here. Let's just go with the preset example. And while that generates, let me show you what we actually did with this, because I thought this one was really impressive because it's not just good at creating, but also editing and repurposing. One of the big new features now is it can output up to 8 different images. So I just went ahead and uploaded a recent YouTube thumbnail. This was the video on the Free Level Framework. I highly recommend watching that if you're a long-term viewer of the channel. I'll be referencing this over and over again. It's really good. It talks about how to get into agents. And I said, create social media graphics from this thumbnail in 3 sizes. Instagram square, Twitter banner, Pinterest pin. And look at that. Instagram square, Twitter banner, Pinterest pin, all from one prompt. I am critical of these things. The composition is good. No, it's not good. It's perfect. Look at it resizing the font in a way that is actually tasteful. To me, this is unbelievable. It does the work of a graphic designer, right? You could now repurpose one thing into many others easily. Beyond that, it has near-perfect text rendering. It can do entire restaurant menus. And it's not like every fifth word has a little mistake. Now it's actually near-perfect. It can output 2K resolution. It can do various formats, as you just saw right there. And they dramatically improved the instruction following, meaning if you tell it to include certain words, it will actually include them rather than take creative freedom with it. And yeah, also, as I showed you, the ability to actually retain your face and like remix yourself into different situations or create professional headshots, that's a really popular use case, is, I would even argue, better than Gemini's Nanobanana Pro. Overall, I'd state that this is probably the best and certainly the most user-friendly and useful image generator out there now. And it's available on all accounts. Go ahead and try it now if you haven't done so yet. And those are the main stories for this week. But there are a few things that popped up on my radar that I wanted to share. Again, as I shared last week, I want to drop this responsibility of covering everything that came out in AI that you could use. In fact, I'll do this Radar segment of like interesting stuff to keep an eye on. And one of them is Cloud Cowork adding dynamic artifacts. And listen to this, if you code up an application, it can become a permanent part of your workspace, as in you can pin it to the side and the application can now, and this is the kicker, take in live data from your connectors. So if you use some application that works with your existing emails and you connect it to Gmail, well, Every time you open the app, it will pull in your newest emails or any other data from another connector. This makes creating custom apps with custom data so much simpler than ever before. It's so fascinating to watch these user-friendly personal agent type applications develop. And as I said, I think OpenAI will be following suit here because they'll need a cowork type application, their own ecosystem. And I think they're gonna evolve ChatGPT into that very soon. Also interestingly, and this is more on the developer side of things, if you're building something or want to use this, There's a Deep Research Max API now from Google Gemini. So if you're familiar, it's basically Deep Research, but it can connect to your data, the web, and generate full reports in just one API call. Just crushes all benchmarks. This seems like the best research tool on the market. And then also, finally, I want to mention Claude Design that came out last week, which was Anthropic's first stab at a design tool. Looks really solid and people are using it. You create marketing materials, pitch decks, app prototypes, and you could even export directly to Canva or PDF. And this makes them the first major AI company to ship a full visual design product inside the chatbot interface. Worth looking into if you make design assets or even into something like email marketing, social media content creation, or similar. Heck, people are even experimenting and making these little PowerPoint-style videos with it. And that's everything I wanted to share with you today. Busy week. We've been preparing and holding the AI Advantage Summit, online conference designed to save you time with AI. And I just want to again highlight that this is really geared that people who are not at level 3 yet, who are not building agents, entrepreneurs, small business owners, or professionals who are non-technical and are figuring all of this AI stuff out. If you're interested in that, it was also restreamed on this channel. You can find it under the live recordings for the next 7 days. And with that being said, my name is Igor Pogany, and I hope you have a wonderful week.