AI News - 14 Dec 2024

A lot is happening in the AI world. It is extremely difficult to followup, learn and implement. Therefore I am planning to write a weekly AI development news summary. This will help me and my readers to understand about the trends and new developement some of these may be interesting to know and someone can pickup to expand as research or as a product, we don’t know.

Here’s a summary of all the AI news mentioned in the transcript:

Google Announcements

1. Gemini 2.0 Foundation Model
   • Released “Gemini 2.0 Flash.”
   • Features include structured output, code execution, function calling, grounding, and real-time voice and screen-sharing interactions.
   • Gemini 2.0 supports image manipulation, such as transforming or blending images.
   • Analyze videos without audio, identifying scenes and key moments.
   • Future feature for geographic exploration and context.
   • Benchmarks
   • Demo - Introducing Gemini 2.0
2. Project Astra
   • Universal AI-powered assistant with advanced vision capabilities, able to recognize objects, read books, and identify environmental details. It is enabled by Gemini 2.0. Multimodel memory and realtime information. It is multilingual and can switch languages on the fly.
   • Future integration with smart glasses for hands-free interaction.
   • Demo - Project Astra
3. Project Mariner
   • Browser agent prototype for automating repetitive tasks, such as gathering contact information from websites.
   • Enhanced web research tool that works from chrome as extension. It can work as an agent and do work on your behalf.
   • Demo - Project Mariner
4. Jules (Developer Assistant)
   • AI assistant for coding tasks and game assistance.
   • Exploring your virtual world in video games.
   • Reason objects in 3D real world aroud us.
   • Demo - Gemini for Games
5. Some Other Testing of Gemini 2.0
   • Google Gemini 2.0 - The multi-agentic era is here
   • Gemini- AI Research Assistan
   • Gemini 2.0 Blog

OpenAI Announcements

1. Sora Turbo
   • Video generation tool capable of creating 20-second videos and blending multiple video concepts.
2. ChatGPT Canvas
   • Now available to all users, featuring Python code execution and an updated user interface for writing and coding tasks.
3. ChatGPT with Siri Integration
   • ChatGPT accessible via Siri on iPhone 16+ and macOS, with features like screen sharing and enhanced intelligence.
4. Advanced Voice Mode
   • Combines vision and voice to describe objects and read text from images in real-time.
5. Santa Claus Interaction
   • Seasonal feature allowing users to interact with Santa via ChatGPT.

Other Announcements

1. Anthropic Claude 3.5 Haiku
   • Faster and cheaper model available for chatbot applications.
2. Grok’s New Image Generator
   • Introduced a new image generation model using autoregressive mixture of experts.
3. MidJourney Patchwork
   • Collaborative canvas for generating and organizing images.
4. Adobe Reflection Removal
   • AI-powered feature to remove reflections from photos taken through glass.
5. YouTube Automatic Dubbing
   • Translate and dub videos into multiple languages to expand audience reach.
6. Cognition Labs’ Devin
   • AI coding assistant priced at $500/month, designed for large codebase management.
7. Meta Quest & Windows Integration
   • Enables virtual desktops and workspaces in the Meta Quest 3 VR headset.
8. Google Android XR
   • Augmented and virtual reality platform competing with Apple Vision Pro.
9. Tesla Optimus Robot Update
   • Progress in humanoid robots learning to walk on uneven terrain.

Miscellaneous

• Hostinger AI Website Builder: Simplified website creation using AI.
• Virtual Reality and XR developments from Meta and Google.
• AI livestreams starting December 16 to explore tools in real-time.