If you got hit like this from behind, would you fall? Unitree G1 won't

DexWrist: Advancing Robotic Manipulation with Human-Like Wrist Dexterity

Researchers at MIT’s Improbable AI Lab have developed the DexWrist, an innovative robotic wrist designed to enhance manipulation in constrained and dynamic environments.

Microsoft’s AI Shows Promise in Complex Medical Diagnoses

Meta’s AI Models Pave the Way for Natural Virtual Interactions

Meta’s Fundamental AI Research (FAIR) team, in collaboration with its Codec Avatars and Core AI labs, has introduced a family of research-focused Dyadic Motion Models that generate realistic facial expressions and body gestures from audio and visual inputs of two-person conversations

Google Unveils 'Doppl': Try On Outfits Virtually with AI

Data And AI Quote of the Week

Creative Commons Unveils CC Signals to Foster Fair AI Data Use

Creative Commons (CC) has launched the CC Signals project, a new initiative aimed at promoting fairness and reciprocity in how data is used to train AI models.

ElevenLabs Launches 11ai: A Voice Assistant Built to Act on Your Workflow

ElevenLabs has released 11ai (alpha), a voice-first assistant designed to demonstrate the capabilities of its conversational AI platform.

Gemini CLI: Google’s Daring Answer to Claude Code and Codex in the AI Coding Arena

Gemini CLI is a free and open-source command-line interface that brings its AI model, Gemini 2.5 Pro, directly into developers’ terminals

Mu in Action: A Lightweight Language Model Enhancing Windows Settings

Microsoft has introduced Mu, a compact on-device language model developed to support real-time interactions on Copilot+ PCs. Mu is the model behind the new agent in Windows Settings

Moonshot AI’s Kimi-Researcher: Advancing Autonomous AI for Research

Moonshot AI introduced Kimi-Researcher, an autonomous AI agent designed to handle complex research tasks through advanced search and reasoning capabilities.

Oakley Meta HSTN Glasses: Style Meets Tech, But at a Cost

The Oakley Meta HSTN glasses launched by Meta in partnership with Oakley aim to blend Oakley’s athletic design with Meta’s AI technology, creating what’s being called “Performance AI glasses” for athletes and fans

Controlling Humanoid Robots with OpenWBT

The OpenWBT project, created by GalaxyGeneralRobotics, lets one person control humanoid robots like the Unitree G1 and H1 using a virtual reality headset (Apple Vision Pro) and joysticks.

OpenAI Introduces Record Mode in ChatGPT for Transcribing and Summarizing Audio Conversations

OpenAI has rolled out a new feature called Record Mode within ChatGPT, enabling users to record and transcribe spoken content

Midjourney Unveils V1 Video Model: A New Step in Creative AI

Midjourney, known for its impressive image generation tools, has just launched V1 of its Video Model, marking its first venture into video creation.

MiniMax-M1: Advancing AI Reasoning with Efficiency and Versatility

MiniMax-M1, developed by MiniMax, is an open-weight, large-scale language model that combines a hybrid-attention architecture with robust reasoning capabilities. Designed for tasks requiring deep analysis and long-context processing, it offers a compelling option for developers and researcher

TikTok Unveils Enhanced Symphony Generative AI Tools at Cannes Lions 2025

2025, TikTok introduced updates to its Symphony suite of generative AI tools, designed to streamline content creation for marketers, brands, and creators

How Claude's Multi-Agent Research System Was Built and What It Reveals About the Future of AI Workflows

Anthropic has offered an in-depth look at the engineering behind its Claude Research feature, a multi-agent system designed to tackle complex, open-ended research tasks by coordinating multiple Claude models, each assigned specific roles.

While OpenAI, Google, and Nvidia dominate the spotlight, none of them could function without ASML.

AI’s future is being printed in a Dutch lab, one nanometer at a time.

Quote of the week

Copilot Vision on Windows with Highlights Now Available in the U.S.

Microsoft announced the U.S. availability of Copilot Vision on Windows with Highlights, a new feature designed to enhance user productivity and interaction with Windows PCs

OpenAI Introduces o3-pro: A Smarter, More Reliable Model for Pro Users

OpenAI announced the launch of o3-pro, a refined version of its highly capable o3 model, now available to Pro and Team users through ChatGPT and the OpenAI API.

Lumos Robotics Introduces Lus2: A Leap in Humanoid Robot Agility

Shenzhen-based Lumos Robotics has unveiled its second-generation humanoid robot, Lus2,

V-JEPA 2: Meta’s New Self-Supervised World Model for Advanced AI Capabilities

Meta has introduced V-JEPA 2, a Video Joint Embedding Predictive Architecture that marks a significant step forward in AI world modeling.

China Temporarily Disables AI Tools During College Entrance Exams to Curb Cheating

China has temporarily restricted the use of several AI tools during its national college entrance examination, the gaokao.