Today in Generative Media

May derides 'arrogant' tech; AI and science photography; Radiance fields at BAFTAs

News and Opinion

• Brian May claims “nobody will be able to afford to make music” if “monstrously arrogant” tech companies are allowed to train AI using copyrighted material (Guitar.com)

• Will AI jeopardize science photography? There’s still time to create an ethical code of conduct (Nature)

• Radiance Fields at the BAFTAs (Radiance Fields)

• How Runway Hopes to Outrun OpenAI, Google in the AI Video Race (The Information)

• YC grad Deepnight nabs $5.5M for AI night vision software that disrupts a multi-billion-dollar industry (TechCrunch)

• Not a Coder? With A.I., Just Having an Idea Can Be Enough. (New York Times)

Software

• Pika 2.2 just dropped Pikaframes: you can transform one image to another, 10s AI video generation, 1080P and more (X)

• Introducing Motion Prediction: A new AI model designed to accurately predict an actor’s movement, even when they’re partially or fully hidden by objects or other actors in a scene. 🎥 (WonderDynamics on LinkedIn)

• Hugging Face launches FastRTC to simplify real-time AI voice and video apps (VentureBeat)

• ElevenLabs’ new speech-to-text model Scribe is here with highest accuracy rate so far (96.7% for English) (VentureBeat)

• Welcome to the new Phi-4 models - Microsoft Phi-4-mini & Phi-4-multimodal (Microsoft)

Research

• Sparse Autoencoders for Scientifically Rigorous Interpretation of Vision Models (project page)

• Does 3D Gaussian Splatting Need Accurate Volumetric Rendering? (arXiv)

• No Parameters, No Problem: 3D Gaussian Splatting without Camera Intrinsics and Extrinsics (arXiv)

• Towards an AI co-scientist (arXiv)