Discover quality AI tools
Bring together the world’s leading artificial intelligence tools to help improve work efficiency
Articles published by AlbertRoberts
Anthropic's Claude AI exposed user chat records via its Share Chat feature, allowing Google to index conversations containing sensitive data like health records and corporate documents. Anthropic stated shares only appear if users actively publicize links. The issue was partially resolved by Monday.
Anthropic's Claude AI exposed user chat records via its Share Chat feature, allowing Google to index conversations containing sensitive data like health records and corporate documents. Anthropic stated shares only appear if users actively publicize links. The issue was partially resolved by Monday.
Qwen AI Glasses claimed the top spot in the domestic AI glasses market across all channels from January to May 2026. Online retail data showed a 30.4% share from March to April, with the flagship model capturing 19.4% in just two months. Key innovation includes hot-swap battery technology on the right temple, addressing battery life concerns amid fierce competition among over 356 brands.
Qwen AI Glasses claimed the top spot in the domestic AI glasses market across all channels from January to May 2026. Online retail data showed a 30.4% share from March to April, with the flagship model capturing 19.4% in just two months. Key innovation includes hot-swap battery technology on the right temple, addressing battery life concerns amid fierce competition among over 356 brands.
Chinese AI firm DeepSeek makes its 75% price cut for flagship model DeepSeek-V4-Pro permanent, following a limited-time discount. Third-party benchmark data ranks it top globally for cost-performance, with operational costs reportedly one-twelfth of OpenAI's GPT-5.5. The company states the model maintains leading technical performance despite the price reduction.
Chinese AI firm DeepSeek makes its 75% price cut for flagship model DeepSeek-V4-Pro permanent, following a limited-time discount. Third-party benchmark data ranks it top globally for cost-performance, with operational costs reportedly one-twelfth of OpenAI's GPT-5.5. The company states the model maintains leading technical performance despite the price reduction.
Researchers from Zhejiang University and Tencent Youtu Lab have developed AdaMARP, a self-adaptive multi-agent framework to enhance AI role-playing. It addresses limitations like poor environmental awareness and rigid storytelling by introducing a "Scene Manager" that directs narrative flow and a four-channel message format for richer interactions. The framework, trained on literary datasets, aims to enable immersive, dynamic scenarios and has been accepted by ACL 2026.
Researchers from Zhejiang University and Tencent Youtu Lab have developed AdaMARP, a self-adaptive multi-agent framework to enhance AI role-playing. It addresses limitations like poor environmental awareness and rigid storytelling by introducing a "Scene Manager" that directs narrative flow and a four-channel message format for richer interactions. The framework, trained on literary datasets, aims to enable immersive, dynamic scenarios and has been accepted by ACL 2026.
Alibaba’s Qwen AI assistant launches AI ride-hailing enabling full trip booking via natural language It integrates personalized stops cross service scheduling with food orders travel booking and navigation This move shifts local service traffic from traditional apps to AI chat windows reflecting Alibaba’s task oriented AI services strategy in China’s competitive AI assistant market
Alibaba’s Qwen AI assistant launches AI ride-hailing enabling full trip booking via natural language It integrates personalized stops cross service scheduling with food orders travel booking and navigation This move shifts local service traffic from traditional apps to AI chat windows reflecting Alibaba’s task oriented AI services strategy in China’s competitive AI assistant market
ByteDance launched its Seedance 2.0 AI video model on Feb 12 2026 featuring a unified multimodal architecture It enhances physics simulation for complex actions supports 15 second multi angle video with stereo audio and allows input from text images audio and video The model also offers video editing and extension tools lowering production barriers in film and e commerce
ByteDance launched its Seedance 2.0 AI video model on Feb 12 2026 featuring a unified multimodal architecture It enhances physics simulation for complex actions supports 15 second multi angle video with stereo audio and allows input from text images audio and video The model also offers video editing and extension tools lowering production barriers in film and e commerce
