• 100 School
  • Posts
  • đź’Ż Is it game over for OpenAI?

đź’Ż Is it game over for OpenAI?

Gemini 3 demos with mixed feelings, when Google gives you bananas you make insane infographics, this week's prompt from NotebookLM and more...

Author

Max Haining
November 23, 2025

Hey, it’s Max 👋

If you’ve been anywhere near AI Twitter this week, you already know… it’s the week of Gemini 3.

Denny from Google’s Gemini team

My entire feed was full of demos, vague posts from Google’s team, and a bunch of people messing around and stumbling into things that looked impressive.

So here’s the clearest picture I can give you broken down into the parts that matter for your work if you don’t have three hours to scroll through AI slop:

Designers are losing their minds (in a good way)

I kept feeling something between wow and hmm that I think a lot of people felt too. Like, this tweet from a designer where he starts by saying the model felt underwhelming. Same old purple, mid, whatever. But then gives it context (real references, code snippets, layouts) and suddenly the model wakes up.

And honestly, that’s exactly how all of this felt lately. These models aren’t “push button, get perfection.” They’re “give me something real and I’ll show you what I can do.”

Why this matters to you (even if you’re not a designer): This is the first time we’re seeing an AI model that understands structure, not just aesthetics. That means PMs, founders, marketers, even ops teams can prototype at the level of someone who has taste without needing the years of training.

If you work with designers: this is your prototype superpower.
If you don’t: this is how you get 80% there without bothering them.

Web Design in Gemini 3

Marta is calling this the best UI design she’s gotten from AI and I think what makes this interesting isn’t how “pretty” it is. It’s the spacing, the type choices, and the hierarchy that makes it look like something an actual product designer would hand off.

Feels early, but this is the closest AI generated UI has come to looking like something you’d trust in production.

But the reactions to Gemini 3 are not all rainbows

Besides the hype, there were a lot of mixed feelings about this launch:

  • Peter loves Gemini 3 still annoyed that it can’t edit Google Docs

  • Khe saying it’s is fast, smart, great at analysis but still not his daily driver

  • Tom calling Gemini 3 and Nano Banana Pro “so overpowered it’s a joke”

My thoughts: Gemini 3 is not the ChatGPT killer. It’s a model that slaps when you give it structure. Which honestly tracks with how most AI tools behave.

When Google gives you bananas 🍌✨

The funniest part is that Google didn’t even hype this one. The internet did. And suddenly everyone from every field (design, education, medicine, product, research) is testing how far this thing goes. Turns out: pretty far.

A model that keeps surprising people

People are using Nano Banana Pro for stuff that normally sits across five different tools. This thread from Min Choi lists a few creative ways to use the model like Geo-specific images and insane infographics. It's chaotic, but in a way that makes you think okay this is more than a cute image model.

One of the most practical use cases

One of the most interesting threads wasn’t from a designer but it was from Derya Unutmaz, a doctor who used Nano Banana Pro to explain diabetes with a clean, accurate infographic and actually readable text đź¤Ż

If you work in health, education, policy, HR, science communication, or any role where you need to break down something complex for people, this is the image-generation model you should try!

Something to try this week âś…

Below is the simplified version of the prompt NotebookLM used for this whimsical infographic that you could use for any other style:

  1. Find an image you liked and ask Gemini to describe it.

  2. Upload your source to NotebookLM (I uploaded the link to the landing page of the 30 Days of AI challenge).

  3. Edit this description to match your style and use it as your Infographics prompt within NotebookLM.

If you end up trying this, reply with what you got!

Before you go ✌️

If you’re reading this, here’s what you should take away:

  • You don’t need to decide whether Google or OpenAI won.

  • And you’re not behind. You’re early.

The real opportunity right now is in learning how to work with these models before they become invisible infrastructure. If you’re up for a challenge that would help you learn that, join waitlist for the next 30 Days of AI round.

Cheers!

Max đź‘‹

P.S. Want to make your team & company AI-first? Let us help here.