Today in Generative Media

Billie Eilish, Pearl Jam vs. AI; Is Suno trained on copyrighted music? Election deepfakes

• Billie Eilish, Pearl Jam, 200 artists say AI poses existential threat to their livelihoods (Ars Technica)

• Suno is a music AI company aiming to generate $120 billion per year. But is it trained on copyrighted recordings? (Music Business Worldwide)

• An A.I. Researcher Takes On Election Deepfakes (New York Times)

• How the Ad Industry Is Making AI Images Look Less Like AI (Wall Street Journal)

• How Good is Sora Actually? An Interview with the Filmmakers Behind "Air Head" (NoFilmSchool)

• Deepfakes Are Coming for the Financial Sector (Wall Street Journal)

• Beyoncé’s new album ‘Cowboy Carter’ is a statement against AI music (TechCrunch)

• A Deepfake Taylor Swift is Teaching Math to Kids on TikTok (PetaPixel)

• Introducing Stable Audio 2.0 – a new model capable of producing high-quality, full tracks with coherent musical structure up to three minutes long at 44.1 kHz stereo from a single prompt. (X)

• This repository contains the PyTorch implementation of Readout Guidance: Learning Control from Diffusion Features. (GitHub) Model weights on HuggingFace.

• Deep Dive into Sora’s Diffusion Transformer (DiT) by Hand (Medium)

• Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic Gaussians (CVPR 2024, project page)

• CameraCtrl: Enabling Camera Control for Text-to-Video Generation (project page)

• CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians (project page)

• Getting it Right: Improving Spatial Consistency in Text-to-Image Models (project page)

• Added an experimental "composition only" transfer to the IPAdapter. (X)