The AI Battle Just Got Real: Why Gemini Has Everyone Worried

Code Red has been called at the Open AI Band Camp ever since Gemini 3 released which came as a surprise. Sam Alman has reportedly called Code Red literally just because of Gemini. That took every I think that took everyone by surprise to be honest. We knew it was coming. We did not know how good it is. And I've been banging on the last couple of weeks about how it's a good enough model for most people. And this is Chachi BT's domain. We were talking about just before the show like they're throwing things at Sora Atlas. They're they're trying to do so much and really it just comes down to when you know they've earned the the noun chat GBT, haven't they? Oh, I'll chat GBT or chat it or whatever. You know, sometimes I say chat GBT even though actually I'm a Ge I'm a Gemini soy boy. Do you know what I mean? So yeah. You say like just ask Chai GPT, but you don't mean just like you know an open AI product. You mean like go use any AI? There's a new model that they're working on that is basically a direct response to to Gemini. Chbt Garlic is the new uh model that they are hopefully going to work on that's going to compete. Speaking of bad names. Yeah, I know. Yeah. But saying all that, you know, I'll say it again. I they're competing in an area where norm they're the normie they're the normally um AI company right and uh Gemini is good enough right so and and people the usage of Gemini has gone up since Gemini 3 I've doubled down on Gemini actually I've always preferred Gemini um and now I'm I literally I think I think perplexity was my action button but now it's is Gemini. Google might be like positioned best out of every other company to win in pretty much every uh in every what do you call it like yeah vertical. So for image they are Google for video they have YouTube you know saying all that I want to I want to ask the audience a question like what AI are you using? Are you are you like a multi-u multi- AI user for different things? Do you find that you have one AI and you just stick to it? Um because I mean, yeah, I'm I'm sort of in that camp, but I use Gemini pretty much now because it's good at everything else. And then Anthropic for code. I might dip into chat GBT for code and not even chat GBT. Sorry. I might chat jump into um GBT 5.1 for like a lot of a heavy thinking task, but not really. I it's just another option that I throw it when I'm coding, but I'd be interested to know what our users are using, whether it's generic. Interesting. Um yeah, let us know even like in the chat here. Just write down Gemini literally just released Gemini 3 deep think yesterday or day before. So, they already kind of in the lead and then they drop this massive very very hardcore model. It's only available on the $250 per month Google AI Ultra plan. Um, and it's achieving like 93.8 on the GPQ is that GPQA diamond and um 45.1 on the RKGI2 with codeexusion. This is a beast model, but this is where we're starting to get into the realm of where I'm kind of leading us down this path here where we're kind of getting into a very specialist model. Like I don't think you or I need Gemini 3 deep think. I think people who are on the cutting edge of research who are scientists, mathematicians like um healthcare like pro it's not an everyday model you know it's it's a very specialist model I think I mean they've they've named it Gemini 3 deep think but it's almost like it deserves its own kind of name I don't know I don't know I'm not not particularly great at branding and names and things like that but the point is like Gemini or Google have leaprogged themselves in the time that it's taken, you know, Open AI to to respond to um to anything that Google have done. So, you know, it's already it's already heating up. Um and then we've got obviously Deepseek 3.2 coming out which is and they released two models special than their normal version which is on par with GPT5 high. So, that's an open-source model. Yes, it's a little bit behind, but it's an open-source model where I don't know, it it's sort of creating a bit of um competition to say the very least, but I think again I use the term splitting hairs. We're just really splitting hairs at this point when it comes to just AI. I think we'll get to a point where open source models will be perfectly fine, which again leaves chat GBT in the in the dust a little bit. Um because why would people pay for something that computers are no doubt going to be good enough in the next 5 years to probably run the intelligence that that open that that these closed source models can only can run right now. You know that'll be available on everyone's machine. So, it's all it's all heating up and um I just wonder whether we'll stop looking for a do it all model and we'll move into a model that can do the thing that we want. Jesus Christ, like six months ago, a year ago, I was handwriting my own code and now I'm deciding whether this AI, this LLM is better than this LLM. It's just it's just ridiculous, isn't it really? I I I feel like this is like screwdrivers for electricians in a way. Like they they could sit and argue which brand is better, but like at the end of the day, the the work is just far beyond what the screwdriver uh alone is. And I think in our work like we'll talk about this in a minute like when we get to anthropic and bun but it's like development and uh software engineering it's just such a big thing and these are just tiny tools still comp comparatively to to the whole vertical. This is part of a larger conversation on my show command AI which we stream live every single week. We discuss the news and all things related to AI in the world of design and web.