Ant Group Launches Cloud-Native Intelligent Capacity Tool Kapacity V0.2 Integrating Self-Developed AI Algorithms

Ant Group has announced the launch of Kapacity V0.2, an intelligent cloud capacity technology tool. This version introduces an AI algorithm developed by Ant Group, known as "traffic-driven replica number prediction," which enables predictive elastic scaling at a production level. Predictive scaling has advantages over reactive scaling as it can anticipate changes in traffic, stabilize resource levels, and utilize resources with higher precision. Kapacity uses a traffic-driven replica number prediction algorithm, transforming the problem of replica number prediction into a traffic prediction problem. By learning the relationships between traffic, capacity, and replica numbers through linear and residual models, it achieves accurate prediction of replica numbers. Additionally, Kapacity offers custom metric support, extending the Metrics API and integrating with Prometheus Adapter to enable historical queries of general metrics and aggregation queries by workload dimensions.

Brazilian Regulators Ban Meta from Using Citizen Data for AI Model Training

The Brazilian National Data Protection Authority (ANPD) issued a "precautionary measure" prohibiting Meta, Inc. from using its citizens' data to train AI algorithms. This decision immediately suspended the changes made to Meta's privacy policy last month, which added the company's AI training to its data handling disclosures.The ANPD stated: "Following an initial analysis, considering the serious and irreparable risks to users, the agency has temporarily ordered the suspension of the privacy pol

Aliyun Open Source Database AI Algorithm PilotScope

Aliyun has launched the free open-source database AI algorithm PilotScope. PilotScope enables one-click deployment of AI algorithms in databases, improving the speed of tasks such as query optimization by 1 to 2 times. PilotScope has been piloted internally at Aliyun and is now freely available as open-source.

U.S. Largest Health Insurance Company Accused of Using AI Algorithm with 90% Error Rate to Deny Claims, Facing Lawsuit Allegations

UnitedHealthcare, the largest health insurance company in the U.S., has been accused of using an AI algorithm with a 90% error rate to bypass doctors' judgment and deny critical medical coverage to elderly patients. The AI algorithm developed by NaviHealth, a subsidiary of UnitedHealth, has led to patients prematurely exiting rehabilitation programs and care facilities, forcing them to deplete their savings. Two family members of denied insurance claim patients have filed a lawsuit in Minnesota District Court, seeking to turn it into a class action that could impact tens of thousands nationwide.

Wenxin Yanyi Monthly Active Users Exceed 200 Million, Baidu Integrates AI Entry Points Across the Entire Ecosystem

The competition in China's generative AI sector has entered a new phase focused on user acquisition. "Baidu Wenxin Yanyi" has exceeded 200 million monthly active users, becoming the first domestic large model to enter the 200 million club. "Alibaba Tongyi Qianwen" achieved over 100 million monthly active users within two months of launch, showing rapid growth. Baidu has adopted an "all-domain integration" strategy, fully embedding AI capabilities into core products such as search, accelerating application implementation.

10B-Parameter Small Nuclear Bomb: Stepwise Star Open-Source Step3-VL-10B Performance Challenges 200B Large Models

The Stepwise Star open-source multimodal vision-language model Step3-VL-10B excels in multiple benchmark tests with only 10B parameters, solving the problem of insufficient intelligence in small models. The model achieves the best performance in its scale in visual perception, logical reasoning, and math competitions, even surpassing open-source and closed-source flagship models that are 10 to 20 times larger in size.