Today in Generative Media

Hannah Diamond & AI music; Adobe trains AI with bananas; AI watermarks defeated

• Hannah Diamond Has Cracked the Code of Using AI for Music (Wired)

• Adobe Wants 1,000 Photos of Bananas to Help Train its AI Image Generator (PetaPixel)

• Researchers show how easy it is to defeat AI watermarks (Ars Technica)

• Runway Partners with Canva (RunwayML)

• Hotshot-XL is an AI text-to-GIF model trained to work alongside Stable Diffusion XL. (GitHub)

• Interactive Scenes built with Gaussian Splatting. Available TODAY in Luma iOS App, Luma Web, as well as the Luma API. (X)

• Fast Stable Diffusion XL on TPU v5e (HuggingFace)

• Announcing torch2jax! Run PyTorch code natively in JAX. 🤝 (X)

• MVDream (Text-to-3D) Colab 🥳 (X)

• ActorsNeRF: Animatable Few-shot Human Rendering with Generalizable NeRFs (ICCV 2023, project page)

• LLM-grounded Video Diffusion Models (project page)

• ImagenHub: Standardizing the evaluation of conditional image generation models (project page)

• Representation Engineering: A Top-Down Approach to AI Transparency (project page)

• We can start to see, dimly, what the near future of AI looks like. (Ethan Mollick)

• facebook messenger has ai generated stickers now and I don't think anyone involved has thought anything through (X) - slightly NSFW!