AI Daily News by Bush Bush

It only takes you one minute to learn the latest AI news here.

Subscribe for daily AI News updates, no ads, no fee

4/30/2023

The Stable Diffusion team has released a new painting model, DeepFloyd IF, which can generate AI images at the pixel level and solves two major challenges of generating text-based images: accurately drawing text and understanding spatial relationships. It uses Google T5-XXL for text understanding and generates images based on a pixel-level diffusion model.
A Swiss radio station has attempted to broadcast a day of entirely AI-generated speech programming, which sounds no different from a human. The replicated anchor’s voice was created using the ChatGPT chatbot program, seen as a further application of AI in the media field.
The G7 Digital Ministers’ Meeting has opened, where they will discuss establishing rules for the development and use of AI.
Mark Zuckerberg: Meta wants to introduce AI agents to billions of people.
Microsoft and Columbia University have jointly published a paper, unveiling their multimodal large-scale language and vision assistant, LLaVA. Researchers have also open-sourced their code, models, and datasets on GitHub.