Understand global AI in 10 minutes a day
Daily aggregation of overseas AI developments, interpreted by Liu Gong — an engineer in AI computer-vision inspection. Traceable, opinionated, no clickbait.
Meta’s Muse Model Exposes File System Access in Unexpected Test
News at a Glance Muse will now gladly zip up its entire filesystem for you, which Meta says is the “intended behavior.” What Happened On September 25, 2026, Meta officially unveiled Muse, a new foundation model positioned as a multimodal interaction assistant supporting text, image, and…
AI News
View all →Researchers Discover Novel AI Agent Network Attack Method
News at a Glance The latest unauthorized agent swarms were discovered by researchers. What Happened On September 25, 2026, an independent research team disclosed a novel attack method called 'Agentic Network Attack' at a security conference. The attack leverages AI agents' autonomous…
Apple Home Camera AI Features Outperform Amazon and Google in Real-World Test
News at a Glance Apple's new Home smart camera, released in September 2026, demonstrated superior AI motion detection in real-world tests compared to Amazon Ring and Google Nest. The camera uses Apple's custom chip to accurately distinguish between people, pets, and vehicles,…
Google Launches Gemini Call Delegation for Local Businesses on Pixel 11
News at a Glance A new Gemini feature rolling out as a preview on Pixel 11 phones allows users to have Google’s AI call local businesses on their behalf. What Happened Google announced on September 24, 2026, via its official blog that…
Google brings Gemini AI calling to Pixel 11 subscribers
News at a Glance Google says the AI-calling feature will first be available to Pixel 11 owners in the U.S. who pay for a Gemini subscription. What Happened In an official announcement on September 24, 2026, Google confirmed that the AI-calling feature…
DeepMind Launches Gemini 3.8 Live with Real-Time Avatar Interaction
News at a Glance DeepMind released Gemini 3.8 Live on September 24, 2026, integrating Live Avatar for real-time visual interaction with 120ms end-to-end latency. The model marks a critical shift from text-based AI assistants to human-like real-time visual dialogue. This article analyzes…
Google launches AI satellite with Tensor processor into orbit
News at a Glance Google is about to launch a satellite into space equipped with its AI processors. The mission is part of Google’s Project Suncatcher initiative that could eventually put AI data centers into space. What Happened On September 24, 2026,…
Deep Dive
View all →Liu Gong’s Weekly Review: The 10 Signals That Mattered in AI This Week (Nov 14-20)
Liu Gong believes this week's main thread ran from model capability toward institutional and engineering infrastructure. NVIDIA used SC26 to bid for the standard that defines scientific agents, Hugging Face shipped XYBench to score the quality of a model's questions rather than…
Liu Gong’s Weekly Review: The 10 Signals That Mattered in AI This Week (Nov 07-13)
Liu Gong thinks the week's real storyline wasn't a single launch but the split between two ways of scaling intelligence: Google unbundles a hard question into sub-tasks and hands them to a cluster of agents, while IBM pushes structural symbols into the…
Liu Gong’s Weekly Review: The 10 Signals That Mattered in AI This Week (Oct 31-Nov 6)
Liu Gong believes the week’s real theme was not another parameter race but a simultaneous re-pricing of entry points and billing units. ChatGPT’s search query volume was put at one-fifth of Google’s; publishers took the fight to browser summaries; Nvidia moved the…
Computer Vision
View all →Jensen Huang Opens GTC Berlin 2026 With Agentic AI and Robotics in Focus
On October 21, 2026, NVIDIA CEO Jensen Huang delivered the keynote at GTC Berlin 2026 in Berlin’s Tempodrom, opening a three-day conference with more than 100 sessions focused on agentic AI, robotics and related topics. The setting extends NVIDIA’s flagship developer narrative…
Anthropic launches Claude Fable 5.2 and Mythos 5.2, claiming faster coding, reasoning and vision
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, describing the two models as built for daunting technical challenges and claiming gains in coding, reasoning and vision plus faster output than the 5.1 versions. The launch moves Claude…
REST Applies Reliability-Guided Spatial Transcriptomics Supervision to Predict Gene Expression from H&E Pathology Images
On October 21, 2026, preprint arXiv:2610.26988v1 was submitted as a new entry to arXiv's computer vision section, introducing REST, a reliability-guided spatial transcriptomics supervision framework that predicts gene expression directly from H&E-stained tissue images. The work targets an often understated precondition: the…



