Today in Generative Media

AI site claims terrorist link; Deepfaked Sam Altman doc; AI risks in game VO proposal

News and Opinion

• Yale Suspends Scholar After A.I.-Powered News Site Accuses Her of Terrorist Link (New York Times)

• This Director Couldn’t Get An Interview With Sam Altman. So He Deepfaked Him In a Documentary (Hollywood Reporter)

• 8 months into their strike, videogame voice actors say the industry's latest proposal is 'filled with alarming loopholes that will leave our members vulnerable to AI abuse' (PC Gamer)

• French publishers and authors file lawsuit against Meta in AI case (Reuters)

• Sakana claims its AI-generated paper passed peer review — but it’s a bit more nuanced than that (TechCrunch)

• Over half of American adults have used an AI chatbot, survey finds (NBC News)

• Is it cheating? AI use during job interviews sparks debate over whether to restrict emerging tools (GeekWire)

• I'm excited to announce that Asteria Film Co has partnered Moonvalley to create Marey. This is a clean AI video model made with ethically sourced, licensed material. (Paul Trillo LinkedIn)

Software

• Experiment with Gemini 2.0 Flash native image generation (Google Developer Blog)

• AI-Powered Virtual Production: Keep the Talent. Control the Environment (voia.com)

• Code release for BiGS: Bidirectional Primitives for Relightable 3D Gaussian Splatting (GitHub)

• Your Pre-Production Friend Powered by AI Made by filmmakers for filmmakers, Filmustage pre-production software uses proprietary AI technology to cut costs, time, and hassle. It's your go-to to simplify everything from script breakdowns and scheduling to analysis, synopsis crafting, and budget management.

• Introducing Gemma 3: The most capable model you can run on a single GPU or TPU (Google Blog)

• OlympicCoder-7B is a code model that achieves strong performance on competitive coding benchmarks such as LiveCodeBench and the 2024 International Olympiad in Informatics. (HuggingFace)

• whoa, Remade AI just dropped 8 open source video LoRA effects for Wan 2.1 on HuggingFace: Squish , Cakefy , Inflate , Deflate , Shooting , Rotate and Muscle (X)

Research

• FoundationStereo: Zero-Shot Stereo Matching (project page)

• Tight Inversion: Image-Conditioned Inversion for Real Image Editing (project page)

• PGSR: Planar-based Gaussian Splatting for Efficient and High-Fidelity Surface Reconstruction (project page)

• 7DGS: Unified Spatial-Temporal-Angular Gaussian Splatting (project page)

• NeuraLoc: Visual Localization in Neural Implicit Map with Dual Complementary Features (project page)

• PCGS: Progressive Compression of 3D Gaussian Splatting (arXiv)

• PE3R: Perception-Efficient 3D Reconstruction (arXiv)

Misc

• we trained a new model that is good at creative writing (not sure yet how/when it will get released). this is the first time i have been really struck by something written by AI; it got the vibe of metafiction so right. (X)

• This is effortless fashion. Made with Pika Labs Pikaswaps feature (X)

• AI phone agent realizes it is talking to a parrot (X)

• “Yo!” “Dude!” “Bro!” (X)