Today in Generative Media

'Hundreds of ads in minutes'; Gannett axes AI review site; AI reassembles ancient epic

News

• ‘Hundreds of Ads in Minutes’: This Startup Thinks AI Can Drown Out the MAGA Movement (Wired)

• Gannett is shuttering site accused of publishing AI product reviews (The Verge)

• Piecing Together an Ancient Epic Was Slow Work. Until A.I. Got Involved. (New York Times)

• Canadian Generative AI Animation Start-Up Viggle AI Raises $19 Million (Cartoon Brew)

• Apple is expected to debut the first generative AI iPhone at its September 9 event (CNN)

• OpenAI supports California AI bill requiring 'watermarking' of synthetic content (Reuters)

• The AI Guys Are Driving Themselves Mad (New York Magazine)

• Hello, you’re here because you said AI image editing was just like Photoshop (The Verge)

• Hobbyists discover how to insert custom fonts into AI-generated images (Ars Technica)

Software

• What the Gaussian Splat? Niantic Scaniverse 4 invites everyone to share their world in 3D (Niantic Labs)

• ComfyUI for Nuke (GitHub) API to be able to use ComfyUI nodes within nuke, only using the ComfyUI server.

• ViTMatte for Nuke (GitHub) This project brings “ViTMatte - Boosting Image Matting with Pretrained Plain Vision Transformers” to The Foundry's Nuke

• CogVideoX-5B, Open weights Text to Video AI model is out Bigger size, better quality, lower cost (runs on just 12GB GPU) (X)

• gsplat 1.3.0 now officially supports 3DGS compression. Easily get 15x compression rate in one line of code (X)

• Google just released official code for RB-Modulation! Want to condition your diffusion model on a subject or style image without using IP-Adapters, then try our method. (works on StableCascade and is adaptable to SDXL and Flux) (X)

Research

• MeTTA: Single-View to 3D Textured Mesh Reconstruction with Test-Time Adaptation (BMVC 2024, project page)

• MLMove Learning to Move Like Professional Counter-Strike Players (SCA 2024, project page)

• SwiftBrush v2: Make Your One-step Diffusion Model Better Than Its Teacher (arXiv)

• GSFusion: Online RGB-D Mapping Where Gaussian Splatting Meets TSDF Fusion (arXiv)

Misc

• Adobe is on the lookout for filmmakers, VFX pros, and motion designers using AI in their work. We’re offering a paid hour for a 1:1 conversation. (X)

• I trained a LoRA (via astria.ai, highly recommended), prompted "a character reference sheet of JONBARRON in a Star Trek cartoon", cropped+centered two faces, and Dream Machine'ed them. Tada, surprisingly stable 3D rotation. (X)

• I first wrote this on the back of a student's paper, and two weeks later he turned in a paper about an issue he really cared about. It earned an A. (X)