My Favorites This Week
ChatGPT Memory Upgrade
Apr 11, 2025
ChatGPT’s memory now references all past chats, enabling more personalized responses based on your preferences—for writing, advice, learning, and more.
Google Veo 2 GA
Apr 10, 2025
Veo 2 is now production-ready in the Gemini API, enabling high-quality video generation from text and image prompts with physics simulation and diverse visual styles.
Google Firebase Studio
Apr 9, 2025
Google launches Firebase Studio, a Gemini AI–powered web platform for building, testing, and deploying apps—streamlining development within the Google ecosystem.
The Firebase Blog Introducing Firebase Studio
Gemini 2.5 Pro Deep Research
Apr 9, 2025
Google’s Deep Research now runs on Gemini 2.5 Pro, delivering powerful data analysis and insights to streamline complex research tasks.
Google Dave Citron Deep Research is now available on Gemini 2.5 Pro Experimental.
Meta AI Llama 4
Apr 6, 2025
Meta launches Llama 4 Scout and Llama 4 Maverick, its most advanced multimodal models yet—outperforming top rivals across benchmarks with unmatched efficiency and visual grounding.
Meta AI The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation
Other Releases
Canva Visual Suite 2.0
Apr 12, 2025
Canva Visual Suite 2.0 introduces AI-driven features that allow users to create everything from documents and presentations to websites—all within a single design.
Pika Twists
Apr 10, 2025
Pika Twists lets you edit any character or object in your video while leaving the rest untouched—giving you precise creative control.
Amazon Nova Sonic
Apr 10, 2025
Amazon’s Nova Sonic is a speech-to-speech AI model enabling real-time, natural conversations with expressive tone, emotion, and seamless speech generation.
Amazon Web Services, Inc. Amazon Nova speech models – AWS
Google Agent2Agent Protocol
Apr 9, 2025
Google’s Agent2Agent (A2A) Protocol enables seamless communication and collaboration between AI agents across vendors and platforms, unlocking secure, cross-agent interoperability.
Google for Developers Blog - News about Web, Mobile, AI and Cloud
Google Agent Development Kit
Apr 9, 2025
Google’s Agent Development Kit (ADK) offers a modular framework for building AI agents with Gemini models, supporting both simple and complex multi-agent applications.
Google for Developers Blog - News about Web, Mobile, AI and Cloud
Google Ironwood TPU
Apr 9, 2025
Google unveils Ironwood, its most powerful and efficient TPU yet—purpose-built for inference and optimized for large-scale “thinking models” like LLMs and MoEs.
Google Amin Vahdat Ironwood: The first Google TPU for the age of inference
NVidia Llama 3.1 Nemotron Ultra 253B
Apr 8, 2025
NVIDIA’s Llama 3.1 Nemotron Ultra 253B is a reasoning-optimized LLM with 128K context length, fine-tuned for chat, RAG, and tool use—runnable on a single 8xH100 node.
NVIDIA NIM llama-3.1-nemotron-ultra-253b-v1 Model by NVIDIA | NVIDIA NIM
👏 Thanks for reading!