Sunday Rundown #120: AI Shopping & AI Comedy

Sunday Bonus #80: Swipe file with 100+ use cases for Nano Banana Pro.

Happy Sunday, friends!

Welcome back to the weekly look at generative AI that covers the following:

• Sunday Rundown (free): this week’s AI news + a fun AI fail.

• Sunday Bonus (paid): an exclusive segment for my paid subscribers.

---

In case you missed it, here’s this week’s Thursday deep dive:

6 Practical Use Cases for Sora 2 Image-to-Video

---

Note: If you’re consistently missing out on my emails, remember to check your “Promotions” tab and mark whytryai@substack.com as a “Safe Sender.”

---

Let’s get to it.

🗞️ AI news

Here are this week’s AI developments.

👩‍💻 AI releases

New stuff you can try right now:

1. Alibaba open-sourced Z-Image, a fast text-to-image model with bilingual text support and excelling at photographic images.

2. Anthropic released Claude Opus 4.5, which tops coding and agentic tool use benchmarks while being more aligned and resistant to prompt injection.

3. Black Forest Labs open-sourced FLUX 2, an image model that writes complex text, maintains character consistency, and outputs in 4MP resolution.

4. Character.AI debuted Stories, letting anyone craft choose-your-own AI adventures featuring their favorite characters.

5. Google news:

   • Agentic Calling lets US customers ask Google to call local businesses on their behalf to gather information about availability, pricing, and more.

   • Nano Banana Pro is now built into Workspace tools like Slides, Vids, and NotebookLM, helping you create infographics, visuals, and slide decks.

6. LTXStudio rolled out Retake, a feature that lets creators rewrite dialogue or tweak a shot in an AI video without having to regenerate it from scratch.

7. Microsoft open-sourced Fara-7B, a compact agentic model tuned for real-world web tasks and computer use.

8. OpenAI news:

   • Improved Voice Interface in ChatGPT means you can now use voice chat while viewing the transcript and visual results in the chat window.

   • Shopping Research crafts a personalized “Buyer’s Guide” from your request directly inside ChatGPT.

9. Perplexity news:

   • Assistants with Memory can remember your preferences across chats and lean on this history to provide more relevant, tailored responses.

   • New Shopping experience can have the AI assistant find the right product for your needs and lets you complete purchases with one click.

10. Tencent open-sourced HunyuanVideo‑1.5, a model that creates high‑quality, smooth videos from text or images.

---

📖 AI resources

Helpful AI tools and stuff that teaches you about AI:

1. “Estimating AI productivity gains from Claude conversations” [STUDY]—Anthropic’s research into measuring Claude’s impact on productivity.

2. “HumaneBench” [REFERENCE]—a benchmark measuring the steerability and “humaneness” of AI models.

3. “The Thinking Game” [DOCUMENTARY]—a great look at Demis Hassabis, DeepMind, and the journey toward building AGI.

🤦‍♂️ AI fail of the week

Sora 2 made this “Funniest video of all time.”
I think human comedians are safe for now.

Send me your AI fail for a chance to be featured in an upcoming Sunday Rundown.

---

💰 Sunday Bonus #80: 100+ use cases & prompts for Nano Banana Pro (swipe file)

Google’s been shipping like crazy lately.

So hot on the heels of last week’s Gemini 3 swipe file, I now also have one for the awesome Nano Banana Pro. Google’s new image model can do crazy stuff like fully fleshed-out infographics with accurate text and much more.

Today’s swipe file adds to this growing list:

• 90+ use cases for GPT-4o image generation

• 45+ practical use cases for o3

• 40+ use cases for Claude 4

• 72 use cases for Nano Banana

• 100+ use cases & prompts for Gemini 3 Pro

There are 103 use cases across 7 categories, with an additional filter to separate image creation from image editing/remixing use cases: