Weekly AI #18

Welcome to Weekly AI #18.
At GTC 2025, NVIDIA unveils its next-gen Blackwell Ultra GPUs and Project GR00T breakthroughs, pushing humanoid robots closer to AGI with real-time physics simulation and embodied AI training at scale.

Top 5 Tools

Remini - Remini's photo and video enhancer helps you improve, sharpen and unblur your media files instantly

WhisperTranscribe - Generating and repurposing content from audio is now quick and easy

Twos PALs - Organize notes, to-dos, reminders & more

Big O Calculator - The time complexity calculator of Big O Calculator makes it easy

SynCity - Training-free generation of 3D worlds

Top News/ Shorts

Mon, Mar 17
- Freepik integrates Gemini for AI-powered image retouching
- Mistral AI releases Small 3.1, a powerful open-source AI model
- Android Studio updates Gemini with multimodal support and Enterprise Tier
- Rendernet makes high-quality product ads effortless with AI
- Gemini 2.0 Flash removes watermarks raising ethical concerns
- EngineAI PM01 performs Axe Gang dance moves from Kung Fu Hustle

Tue, Mar 18
- Nvidia unveils Blackwell Ultra, the successor to Blackwell AI chips
- NVIDIA launches GR00T N1, an open-source humanoid robot foundation model
- Jensen Huang introduces Blue droid after NVIDIA's DeepMind and Disney partnership
- NVIDIA and Google DeepMind launch Newton, an open-source physics engine for robots
- NVIDIA CEO introduces DGX Spark, the world's smallest AI supercomputer
- Windsurf AI launches Wave 5 with faster predictive Tab and free access
- Nvidia and GM partner to advance self-driving car technology
- Adobe launches AI agents for enhanced marketing and customer experience
- Stability AI launches Stable Virtual Camera for 3D video creation
- Google Gemini introduces Canvas and Audio Overview for enhanced productivity
- Kling AI introduces DizzyDizzy Effect for immersive video fun
- Claude is down but status page shows no reported issues
- Aha AI platform revolutionizes influencer marketing with automation
- Sync creates the most natural lipsync model in the world

Wed, Mar 19
- Orpheus 3B offers high-quality, emotive text-to-speech with Apache 2.0 license
- AudioX introduces anything-to-audio generation with impressive results
- xAI launches Grok Image Generation API for developers
- NVIDIA achieves world's fastest DeepSeek-R1 inference with Blackwell GPUs
- NVIDIA and GE HealthCare advance autonomous diagnostic imaging with Isaac for Healthcare
- Unitree G1 levels up again, it can flip with somersault
- NEO Gamma robot vacuuming the floor at NVIDIA GTC

Thu, Mar 20
- OpenAI launches advanced audio models for text-to-speech and speech-to-text
- Vercel and xAI collaborate to provide Grok models seamlessly
- Claude AI introduces web search with inline citations for paid users
- Apple restructures AI division, assigns Siri to Vision Pro chief Mike Rockwell
- Pika Labs offers exclusive early access to video manipulation tools
- Topaz Labs launches Gigapixel v8.3.0 with Recover v2 for image restoration
- Krea AI introduces Video Training for custom AI video generation
- Google NotebookLM adds mind maps and multilingual content generation
- OpenAI launches o1-pro API with enhanced capabilities and higher costs

Fri, Mar 21
- WPP partners with NVIDIA, Boston Dynamics, and Canon for robotic cinematography
- Google AI explores neural activity alignment with LLM embeddings in human brain
- Anthropic updates Claude Code with extended thinking and improved tool use
- Hunyuan introduces T1, a breakthrough in AI reasoning with Hybrid-Mamba-Transformer
- xAI's Grok introduces image editing with descriptive prompts
- SynCity creates 3D worlds from text prompts with no training required
- Unitree's G1 humanoid robot achieves world's first kip-up
- Vidu AI updates with multi-reference consistency for stable video generation
- ByteDance launches InfiniteYou on Hugging Face for photo recrafting

Sat, Mar 22
- ChatGPT offers unlimited Sora generations for paid users
- Explore fun examples of Co-draw with Gemini 2.0 image generation

Sun, Mar 23
- Apple Watch to feature cameras for enhanced AI capabilities