This weekend in Generative Media

AI assistants leak secrets; AI essays mimic privileged males; OpenAI's nonprofit status

News and Opinion

• AI assistants are blabbing our embarrassing work secrets (Washington Post)

• AI Admissions Essays Align with Privileged Male Writing Patterns (Neuroscience News)

• Fourth Letter to California Attorney General Rob Bonta: OpenAI is No Longer Operating as a Nonprofit (Public Citizen)

• Google Gemini's AI video summary implies YouTube doesn't care about content creators (Android Police)

Software

• Adobe Photoshop Elements and Premiere Elements updated with new AI features (Engadget)

• Google Photos starts rolling out AI-powered ‘Ask Photos’ following waitlist sign-ups (9to5Google)

• BlinkShot: An open source real-time AI image generator. Powered by Flux through Together.ai. (GitHub) blinkshot.io

• Voice-Pro is an integrated solution for subtitles, translation, and TTS (GitHub)

• gradio-client-lite: A zero-dependency, platform-independent, and lightweight Gradio client. (GitHub)

• What?! FAL gives access to Kling, Runway Gen3 and Luma Dream Machine API?! Billed per use?! All in one place?! And you can use them in comfy-style workflows?! Why nobody ever told me?! (X)

Research

• Cafca: High-quality Novel View Synthesis of Expressive Faces from Casual Few-shot Captures (SIGGRAPH Asia 2024, project page)

• ControlNet-XS (project page)

• GI-GS: Global Illumination Decomposition on Gaussian Splatting for Inverse Rendering (project page)

• Disco4D: Disentangled 4D Human Generation and Animation from a Single Image (project page)

• Commonsense-T2I Challenge: Can Text-to-Image Generation Models Understand Commonsense? (project page)

• Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models (arXiv)

• Contrastive Localized Language-Image Pre-Training (arXiv)

• SuperGS: Super-Resolution 3D Gaussian Splatting via Latent Feature Field and Gradient-guided Splitting (arXiv)

Misc

• Very useful tool - Calculates how much GPU memory you need and how much token/s you can get for any LLM & GPU/CPU. (X)

• This could be a piece of fried chicken, or wait, maybe it's actually a little puppy! (X)

• Looks like the new generation of students is better prepared for the age of Gemini/ChatGPT based review... (X)