This weekend in Generative Media

Writers demand legal action vs. AI; AI award controversy; Models use AI DMs

News and Opinion

• Writers Guild Calls on Studios to Take “Immediate Legal Action” Against AI Companies (Hollywood Reporter)

• Former ByteDance Intern Accused of Sabotage Among Winners of Prestigious AI Award (Wired)

• OnlyFans Models Are Using AI Impersonators to Keep Up With Their DMs (Wired)

• AI entertainment studios: The new breed of companies pioneering Gen AI-powered production (Variety)

• Midjourney is launching a multiplayer collaborative worldbuilding tool called ‘Patchwork’ (VentureBeat)

• Character.AI has retrained its chatbots to stop chatting up teens (The Verge)

• How Claude Became Tech Insiders’ Chatbot of Choice (New York Times)

Software and Hardware

• Adobe now has a tool to get rid of ugly window reflections in photos (The Verge)

• Today we launched our Pika 2.0 model. Superior text alignment. Stunning visuals. And “Scene Ingredients” that allow you to upload images of yourself, people, places, and things (X)

• Google launched Gemini 2.0, its new AI model for practically everything (The Verge)

  • Google’s new Trillium AI chip delivers 4x speed and powers Gemini 2.0 (VentureBeat)

  • This repository contains a react-based starter app for using the [Gemini 2.0] Multimodal Live API over a websocket. (GitHub)

• NotebookLM gets a new look, audio interactivity and a premium version (Google blog)

  • Google debuts NotebookLM for enterprises (TechCrunch)

• ChatGPT now understands real-time video, seven months after OpenAI first demoed it (TechCrunch)

• ChatGPT gets screensharing and real-time video analysis, rivaling Gemini 2 (VentureBeat)

• MMAudio — Video-to-Audio Synthesis (HuggingFace)

• Leffa: Learning Flow Fields in Attention for Controllable Person Image Generation (HuggingFace)

• 🔥FireFlow: Fast Inversion of Rectified Flow for Image Semantic Editing (HuggingFace)

• luma / ray: Fast, high quality text-to-video and image-to-video (Also known as Dream Machine) (Replicate)

Research

• Representing Long Volumetric Video with Temporal Gaussian Hierarchy (project page)

• Meshtron: High-Fidelity, Artist-Like 3D Mesh Generation at Scale (project page)

• From an Image to a Scene: Learning to Imagine the World from a Million 360° Videos (project page)

• LiftImage3D: Lifting Any Single Image to 3D Gaussians with Video Generation Priors (project page)

• Using Diffusion Priors for Video Amodal Segmentation (project page)

• Proc-GS: Procedural Building Generation for City Assembly with 3D Gaussians (project page)

• StableAnimator: High-Quality Identity-Preserving Human Image Animation (project page)