Today in Generative Media

OpenAI's 12 days of 'shipmas'; DeepMind's Genie 2: interactive worlds; Weird AI ads

News and Opinion

• OpenAI’s 12 days of ‘shipmas’ include Sora and new reasoning model (The Verge)

• DeepMind’s Genie 2 can generate interactive worlds that look like video games (TechCrunch)

• The abject weirdness of AI ads (TechCrunch)

• Google Introduces A.I. Agent That Aces 15-Day Weather Forecasts (New York Times)

Software

• Veo and Imagen 3: Announcing new video and image generation models on Vertex AI (Google Cloud blog)

• Luma AI in Amazon Bedrock (coming soon): Generate high-quality video clips from text and images, creating captivating motion graphics from static concepts (Amazon)

• Efficient Track Anything (GitHub)

Research

• Motion Prompting: Controlling Video Generation with Motion Trajectories (project page)

• LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting (project page)

• Feed-Forward Bullet-Time Reconstruction of Dynamic Scenes from Monocular Videos (project page)

• One Shot, One Talk: Whole-body Talking Avatar from a Single Image (project page)

• AniGS: Animatable Gaussian Avatar from a Single Image with Inconsistent Gaussian Reconstruction (project page)

• TimeWalker: Personalized Neural Space for Lifelong Head Avatars (project page)

• VideoGen-of-Thought: A Collaborative Framework for Multi-Shot Video Generation (project page)