One's for cinematic ads, the other's for memes. Veo 3.1 vs Sora 2 explained

Hey there, creative human! 🌟 

Google just dropped Veo 3.1, and now it's going head-to-head with OpenAI in AI video. They're playing completely different games though. I'm breaking down which model you should use for what.

Also in this issue:

• Pinterest fights 'AI Slop'

• A prompt for cinematic product photography

• A tool that turns excel data into motion graphics

• The first AI-generated TV commercial ($2,000 production)

Read time: 4 minutes.

Let's dive in! 🚀

STORY OF THE WEEK

📽️ Google VEO 3.1 Launch

Right when Sora 2 videos are flooding the internet, Google drops Veo 3.1. The message is clear: while Sora targets casual creators and viral content, Veo 3.1 is built for professional film and ad production. "Cinematic quality" is their exact positioning.

The update is integrated into Flow (Google's AI filmmaking tool), and here's what it unlocks (in the order of priority as I see it):

1) Ingredients to Video – Upload multiple reference images to control characters, objects, and style:

2) Audio everywhere – All features now include generated audio, not just text-to-video prompts.

3) Frames to Video – Provide a starting and ending frame, and Flow generates the seamless transition between them:

Veo 3.1 also provides Extend to create videos over a minute long by building from the final second of each clip, and Insert to add new elements with automatic lighting and shadow adjustments. A Remove feature is coming soon.

➡️ Use Veo 3.1 in Flow, on Higgsfield, or on Veo3gen.app (offers 75% cheaper prices than the official).

My take on Veo 3.1 vs. Sora 2

Both are impressive, but they're solving different problems.

When you compare the same prompt across both models, Veo 3.1 delivers superior photorealism and that "real film production" quality. Sora excels at motion fluidity and storytelling.

For your workflow? Use both.

If you're making social content, memes, or testing concepts – Sora's speed and ease win.

If you're producing ads, client work, or anything where realism matters, Veo's photorealism and the ability to use multiple input images via "Ingredients" is a game-changer for brand consistency.

🎨 Pinterest Fights ‘AI Slop’

Pinterest has introduced controls that let users limit the amount of generative AI content they see in their feeds across popular categories like beauty, art, fashion, and home décor. This move responds to user backlash against the massive increase in AI-generated imagery and is crucial for maintaining the visibility of human creativity.

This is bigger than Pinterest. It's the first major platform admitting that AI content overload is driving users away.

PROMPT LIBRARY

📦 A Prompt for Stunning Product Photography

This prompt delivers cinematic, professional-grade product images without the photoshoot. Works with ChatGPT or Nano-Banana.

Just replace [product_image], [product_name], and [brand_colors] in the template provided by AltSocietyAI (open the link to copy it).

TOOL SPOTLIGHT

🛠️ Your Must-Have AI

INSPIRED BY AI

🎨 Creative of the Week

Kalshi, the U.S.-based prediction market platform, officially aired this AI commercial during the 2025 NBA Finals broadcast.

The ad, directed by AI filmmaker PJ Accetturo, was produced entirely using Google’s Veo 3 and Gemini (Google’s AI suite). It cost around $2,000 to make, required only two days of production, and became the first nationally broadcast, fully AI-generated TV commercial.

Catch you next week!

— Radu