The Biggest Week in AI Since ChatGPT

Here, have some jet fuel

Your newsletter, now exiting the stratosphere.

We’re beehiiv, the newsletter platform built for builders. Every product we drop and feature we create is made with one goal in mind: exponential growth. 

Our founders sought to bring the tools and tech to creators everywhere that helped turn their old company into a multi-million dollar operation. So when we say we can take you from “up-and-comer” to household name, we mean it.

If you need a beautiful destination for your content, use our no-code website builder. If you want paid subs, we don’t take a dime. If you want big name sponsors, we connect you with the likes of Nike and Netflix. 

In short: maybe your newsletter is stuck on the ground floor because your newsletter platform is stuck in the last decade. 

Level up and get started for free with beehiiv.

The Biggest Week

Phew, this has been an insane week in AI. But if you ask a normal person, they’ll probably have no idea that earth-shaking things just happened.

Why?

All the tools that came out this week are squarely aimed at creators. There wasn’t a ChatGPT moment that registered for most “civilians,” (although GPT-4o’s Image Generation and its Studio Ghibli memes are getting close).

But if you use AI tools daily, your arsenal just got a big upgrade.

I ran through all of the changes in one whirlwind take at the end of the news segment on Niche Pursuits this morning (we also analyzed the impact of Google’s AI push and their March Core Update):

This week, we got a hilarious and very impactful demo from OpenAI of the new GPT-4o Image Generation:

I’m just scratching the surface of its capabilities, but already I’ve found one killer app for creators:

Interestingly, the system also excels at creating boring images. AI often struggles with this. For creators, that’s important, since you don’t always want your illustrations and images to be glitzy, over saturated, and to have that “AI look”.

Here’s my video on making “boring” everyday images with the system:

We also got Ideogram 3, which isn’t as good at GPT-4o at prompt understanding, but makes beautiful images and is great at layout and graphics design. I’m using it for thumbnails:

Finally, Google released Gemini 2.5 Pro. That one is brand new, but it’s already proving great at handling business data.

Oh, and I think Deepseek did something. IDK, maybe I’ll cover it. I don’t really believe in DeepSeek…

Now all we need is a big update to video generators, and our toolkit will be complete. Mark my words, OpenAI has that ready to go, but they won’t release it for a bit, because demand for GPT-4o Image Generation is so high that Sam Altman said on X their “GPUs are melting”.

Probably not literally.

But just in case, here’s a photo from GPT-4o Image Generation of Sam Altman holding a melting NVIDIA GPU (in the style of Studio Ghibli, of course).