Gemini 4 Argon: Google’s AI Comeback

Julian Goldie (0:00)

Gemini 4, Argon just dropped. How good is it? We're gonna come on to that in a second. I'll show you some of the stuff that we've built with it so far. It is capable of building really beautiful websites, as you can see right here. So it can create like really amazing 3D sites. It's also crushing it on benchmarks, although I wouldn't pay much attention to that. And I'll come on to what I think is better instead. And also if we have a look at this here, we've built out tons of stuff on this today, and it can build some really like beautiful visual things, as you can see.

Now, we've already built it into our Gentic Operating System, which I think is the best way to use it. So you can see we've got anti-gravity plugged into this system right here. We can go back and forth of it. We can switch between models. So we've got Gemini 4 Argon. We've got 3.8 Flash, Claude Opus, et cetera. And then inside the build section, we can preview everything that we've created. Now, if we have a look at this, it's capable of creating like amazing designs, 3D worlds, et cetera. It can create like these beautiful 3D animations, like you can see. And the color and the texture on this is super nice as well.

Now, you might be asking, okay, you know, just tell me the truth, Julian.

Is this going to replace Opus 5.5? For me personally, I would say no. I'm just going to tell you that one minute in. And the reason for that is because it's good. It can build some interesting stuff. It can create like these interesting designs and stuff like that. But it's like for design, it's nowhere near the same level. Nowhere near the same level as Opus 5.5. And that's coming from someone who's tested both of them for like dozens of hours at this point. So when we're looking at the system, what I would recommend is that you have a system where you can use the best of everything. So for example, if you got Claude, well, you can use this Claude workflow over here, and you can use it for editing videos, creating content, posting directly to your channels, and analyzing your content performance like we've got over here. And then we can switch over to Antigravity if we need something cheaper, or if we need to use two different agents at the same time. Let's say, for example, you run out of tokens of Claude. No problem. You switch over to Antigravity. Or let's say, for example, you already subscribed to Google Gemini Pro and the AI Pro subscription, and you want to just use the power of it. Well, this is probably the best way to do that, because if you're going directly inside Gemini here, it's not that fun to use, it's not that fun to play with, and also you're going to get limited. And you can see even the default here is switched to 3.8 Flash. Whereas we want to use Gemini for Argo. Whereas if you've got an agentic operating system like this, you know, have a look here. Where's all your builds? Where's all the stuff that you've created? Where's everything in one place? It's not. And how do you switch the agents? Well, you have to go to Claude, and then you have to go to Grockbot, and then you have to go to Obsidian, right? Super messy. Whereas if you've got a system like this, this is way more powerful and fun. And then when you do actually build something good like this or this, for example, I'd say those are two of the strongest builds that were created with it. Well, it's saved inside your workspace, and you come back to it anytime. You never lose anything. Whereas if you're using it inside Gemini, it's messy. It's dated. It feels like you can't really find anything you've created.

I think this is much better. And also, you get the genetic harness from anti-gravity, which means you can switch models at anytime. You can even use Claude Sonnet inside anti-gravity. So if you're coding on something, you can have them working inside one place. The other thing I would say is for 99% of people watching this, like you haven't been using Google Gemini, not day to day, not like Claude or Chad Chibati or Grog Bar. And the reason I say that is because I know what's been going on in AI over the last 12 months. I know that people feel that Google Gemini was falling behind. And so how can you solve that problem? Well, the way that I've solved that, because, for example, if you start using Google Today, it's not gonna have any memory, not gonna have any details, it's not gonna know you like Claude or Chad Chibati that you use all day. And so a big problem is context, because if you don't get a personalized output from the AI, if it doesn't know what projects you're working on, it's a waste of time, right? It's just like having an employee that you have to train for three months before it's any good. So the system that we've got built into this, this agentic OS, as you can see, is that we have the memory galaxy. And that means that all of our memories are stored, but also shared across every AI agent, and it's all automated. So my agents automate this. I don't touch it, right? But every single day, Claude will write some information about me, who I am, what I do, etc. And then, for example, if I fancy using DeepSeek, what I already knows is everything that I've done inside Claude and Hermes automatically. Don't have to mess around. And so I think that's another great way to use and get the most out of anti-gravity, because then you solve the context problem, which is massive, right? There's going to be a lot of people who just say, I'm not going to use Gemini 4 Argon, because it doesn't know anything about me, because it doesn't give me personalized responses, because it's not going to be smart enough when I'm interacting with it. Whereas, for example, if you were going into Claude or Chat Cheapity, it knows everything about you. It has a full memory and everything else. So that's the way that I see this. And then also, the other thing I would say here is what's quite good is you can build something interesting with Gemini 4 Argon, and then you can go back to Claude and you can give it that project detail. So I can say, okay, improve this or deploy this or that sort of thing. I could have Claude and Anti Gravity in the same system, working on the same project. Now also, over here, we have the agent mastermind. So this is our AI agents talking together. And so we can have Hermes working with Gemini and Claude inside the group chat. You can see an example right here. So we can say, hey, Hermes come up with a new idea for an app, and all our agents just jump in and help each other as much as they can. So you can see Google Gemini, and you can use Gemini 4 Argon inside the chat as well, which again, I think is another powerful way of using it.