Today in Generative Media

Silverman vs. OpenAI trimmed; Open source AI narrows gap; Consumers reject AI

News

• Sarah Silverman Lawsuit Against OpenAI Suffers Setback As Judge Trims Case (Hollywood Reporter)

• Open-source AI narrows gap with proprietary leaders, new benchmark reveals (VentureBeat)

• Study finds consumers are actively turned off by products that use AI (Futurism)

• ‘One of the most disgusting meals I’ve ever eaten’: AI recipes tested (The Guardian)

• The EU’s AI Act is now in force (TechCrunch)

• Meta’s advertising growth is proof that hefty AI spending is already paying off (CNBC)

• Zuckerberg touts Meta’s latest video vision AI with Nvidia CEO Jensen Huang (TechCrunch)

• We pushed this ChatGPT game to the limits, but playing it the right way is more fun (Polygon)

• Former GTA dev says 'it's time for a revolution' where 'animation is more AI-driven and physics-driven' than done by hand (PC Gamer)

• This Week in AI: Companies are growing skeptical of AI’s ROI (TechCrunch)

Software

• Announcing Black Forest Labs. Flux.1 Pro: State-of-the-art image generation with top of the line prompt following, visual quality, image detail and output diversity (Replicate)

• Introducing Stable Fast 3D: Rapid 3D Asset Generation From Single Images (Stability blog)

• ChatGPT Advanced Voice Mode impresses testers with sound effects, catching its breath (Ars Technica)

• Nvidia’s coveted H100 GPUs will be available on-demand through Lambda’s clusters (VentureBeat)

• Gemini in the side panel of Google Drive introduces a new PDF viewing experience (Google blog)

• This repository is the official implementation of InstantSplat, an sparse-view, SfM-free framework for large-scale scene reconstruction method using Gaussian Splatting. InstantSplat supports 3D-GS, 2D-GS, and Mip-Splatting. (GitHub)

• Object Eraser Powered by Refiners (HuggingFace). Erase any object from your image just by naming it — no manual work required! Not only will the object disappear, but so will its effects on the scene, like shadows or reflections.

Research

• IntrinsicDiffusion: Joint Intrinsic Layers from Latent Diffusion Models (SIGGRAPH 2024, project page)

• Dynamic 3D Gaussian Prediction for Motion Extrapolation and Free View Synthesis (SIGGRAPH 2024, project page)

• ST-4DGS: Spatial-Temporally Consistent 4D Gaussian Splatting for Efficient Dynamic Scene Rendering (SIGGRAPH 2024, ACM)

• Scale-Invariant Monocular Depth Estimation via SSI Depth (SIGGRAPH 2024, project page)

• Expressive Whole-Body 3D Gaussian Avatar (ECCV 2024, project page)

• Tora: Trajectory-oriented Diffusion Transformer for Video Generation (project page)

• Wolf: Captioning Everything with a World Summarization Framework (project page)

Misc

• The community is at a crisis of copyright and my experience and view of how things are currently going isn’t sustainable. (X)

• Yes, but only one has “open” in their name. (X)

• Rage against the machine learning (X)