Today in Generative Media

California bar used AI; Duolingo subs contractors with AI; AI wrote philosopher's book

News and Opinion

• California Admits AI Was Used to Write Bar Exam Plagued by Problems (Futurism)

• Duolingo will replace contract workers with AI (The Verge)

• A Philosopher Released an Acclaimed Book About Digital Manipulation. The Author Ended Up Being AI (Wired)

• Microsoft’s AI Secretly Copies Your WhatsApp, Signal Messages (Forbes)

• ‘Squared blunder’: Google engineer withdraws preprint after getting called out for using AI (Retraction Watch)

• The MyPillow guy’s lawyers used AI in court and you’ll never guess how that turned out (Futurism)

• Is Anything Real Anymore? AI Making Americans Suspicious Of Everything Online (StudyFinds)

• U.S. government agency sounds alarm on AI's toll on environment, humanity (Mashable)

• 2025 media and entertainment outlook (Deloitte)

Software

• Fast 3D shape prototyping with simple scribble and text prompt. Presented by Tripo. (HuggingFace)

• FantasyTalking: Realistic Talking Portrait Generation via Coherent Motion Synthesis (GitHub) [Now merged in ComfyUI-Wan]

• We present Kimi-Audio, an open-source audio foundation model excelling in audio understanding, generation, and conversation. (HuggingFace)

Research

• Any6D: Model-free 6D Pose Estimation of Novel Objects (project page)

• LoftUp: Learning a Coordinate-Based Feature Upsampler for Vision Foundation Models (project page)

• SmallGS: Gaussian Splatting Based Camera Pose Estimation for Small-Baseline Videos (project page)

• Seurat: From Moving Points to Depth (project page)

• DiTPainter: Efficient Video Inpainting with Diffusion Transformers (arXiv)

• Survey of Video Diffusion Models: Foundations, Implementations, and Applications (arXiv)

Misc

• Here's my 3DV talk (X) [Radiance Fields and the Future of Generative Media, by Jon Barron]