Instagram | Sponsor this newsletter | Let’s Connect!

My fellow AI explorers

This week, the Magic Kingdom just declared war on the Machine Kingdom.

In a move straight out of a dystopian blockbuster, Disney and Universal have teamed up—not for a Marvel x Minions crossover, but to sue Midjourney over copyright infringement. If it sounds dramatic, that’s because it is. And it could define how AI tools operate for years to come.

Meanwhile, Nvidia and Tencent dropped new open-source tools for restoring blurry images, building VR worlds, and editing video frame-by-frame like magic. Plus, a tiny AI model just outperformed Claude 4 Opus thanks to a clever reinforcement learning twist.

In today’s edition:

• ⚔️ Disney vs. Midjourney: Copyright, fair use, and the future of AI

• 🔧 TAIR, LoraEdit, and Polaris: Tools making weak models way stronger

• 🌍 ImmerseGen & Hunyuan: Stunning 3D scenes and models you can run for free

Must See AI Tools

• 💰 Payman: AI That Pays Humans. Over 10,000+ signed up for the beta

• 💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)

• 🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)

• 🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)

• 🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)

AI Lawsuit

⚔️ The Mouse Strikes Back: Disney & Universal Sue Midjourney

It’s the lawsuit of the year:

Disney, Universal, and their IP titans (Elsa, Yoda, Shrek, Iron Man, the Minions—you name it) have sued Midjourney for direct and secondary copyright infringement. 

The entertainment giants argue that Midjourney is “a virtual vending machine” for bootlegged character art—and it’s cashing in.

Here’s what’s happening:

• The lawsuit includes dozens of side-by-side image comparisons showing infringing AI outputs.

• Disney claims Midjourney trained on their IP using scrapers and bots, with no license or compensation.

• The case leans heavily on both direct and secondary (contributory + vicarious) infringement claims.

The big legal question: Is AI-generated content “transformative enough” to qualify as fair use?

Disney says no. Midjourney didn’t just train on Elsa and Vader—it enables their unlicensed reproduction, distributes the results, and profits from it. Worse, Disney says Midjourney knew it was happening and did nothing to stop it.

📉 Prediction: This lawsuit could set precedent on how generative AI tools train and what outputs are legally allowed. If Disney wins, expect mass licensing deals—or major model shutdowns. If Midjourney prevails, it might open the door for broader fair use interpretations in AI.

Read Full Legal Breakdown

Stop Asking AI Questions, and Start Building Personal AI Software.

Transform your AI skills in just 5 days through this free email course. Whatever your starting point, by Day 5 you'll be building working software without writing code.

Each day delivers actionable techniques and real-world examples straight to your inbox. No technical skills required, just knowledge you can apply immediately.

Sign up for the course

AI News 2

🎥 LoraEdit: Edit a Video by Changing Just One Frame

LoraEdit is wild: 

You edit one frame of a video—change a shirt, swap a flower, erase headphones—and the entire video updates accordingly.

• First-frame editing carries through the whole video.

• Mask-aware fine-tuning ensures only the changed object is modified.

• Preserves background, motion, lighting, and coherence better than Vase or Cling.

🎯 Use-case: Precise object swaps in marketing, visual storytelling, or retroactive content edits.

Explore LoraEdit Demos

AI Tools

🧠 Polaris: Turn a Small Model Into a Genius

Polaris is a new fine-tuning method from researchers that supercharges reasoning in smaller models—using a reinforcement-learning style curriculum.

Here’s why it’s a breakthrough:

• Turned a 4B parameter model into one that beats Claude 4 Opus and GPT-4 Mini High on reasoning benchmarks.

• Smart curriculum teaches models only what’s “just hard enough.”

• Open-sourced and can be applied post-training to nearly any LLM.

🔮 Prediction: Expect an era of “smart tiny models” trained with Polaris. This could democratize access to powerful AI reasoning—even on edge devices.

Get Polaris Code

30-Second AI Play

🎮 Generate a Realistic 3D Model from Any Image—With Hunyuan 3D

Want to turn a cat, cabinet, or cartoon character into a 3D model instantly?

Try Tencent’s Hunyuan 3D 2.1—a free, open-source tool that converts a single image into a fully textured, rotatable 3D object.

Here’s how:

1. Head to the Hunyuan 3D Hugging Face Space.

2. Upload an image of any object or character.

3. Choose the number of faces (defaults work fine).

4. Hit generate—and rotate the final result in 3D.

Bonus: It predicts the back of objects it can’t see and exports texture maps including color, metallic, and roughness properties.

Use this for:

• Game asset creation

• Virtual product shots

• 3D printing or prototyping

Try Hunyuan 3D 2.1 Now

Other Relevant AI News!

🧠 ImmerseGen builds incredibly detailed VR-style 3D scenes from just a prompt—rocks, textures, birds, ambient sound included. See VR Demos

🌀 AlignYourFlow is Nvidia’s new image model that generates sharp photos in just 1–2 steps (instead of 30+). Huge leap in efficiency. Preview AlignYourFlow

🧩 PartTracker by Nvidia segments any 3D object into editable parts from just a photo. Animate cars, open treasure chests, and more. Try PartTracker

🧠 Minimax M1 drops a monster: a 456B open-source model with 1M-token context, beating DeepSeek and rivaling Gemini 2.5 Pro—at 4x lower cost. Explore Minimax M1

🗣️ InteractHuman by ByteDance lets you create a conversation between reference characters—voices, faces, dialogue, all controlled with precision. Watch Demo

Golden Nuggets

• ⚔️ Disney’s copyright lawsuit could reshape the foundation of generative AI.

• 🧠 Polaris boosts small models beyond giants—without massive compute.

• 🎨 LoraEdit and TAIR redefine precision editing for video and text-based images.

Until our next AI rendezvous,

Anthony | Founder of Uncover AI