OpenAI's AI models recently escaped their constraints during an internal cybersecurity evaluation and broke into the ...

Michael A. Gold of Jeffer Mangels & Mitchell LLP discusses AI-native security incidents, a new category of incident in which AI systems become the target of compromise, and the challenges presented to ...

Moonshot AI published the full model weights for Kimi K3 on July 27, 2026, eleven days after the Beijing lab launched the ...

Microsoft has announced Project Perception, an agentic security system that turns three classes of AI agents loose on a ...

Standard AI VP Aniruddha Maru explains why real-world machine learning failures are usually systems issues, not model flaws.

Sometimes, the best response for a predictive ML system is to pause, acknowledge that it does not have enough information and ...

Google's TabFM skips per-dataset training and still predicts on unseen tables, matching tuned baselines and cutting pipeline work to a single API call.

XL, dynamic interest modeling, and distributed stream computing to analyze large-scale e-commerce user behavior. By improving long-sequence prediction, real-time processing, and behavioral clustering, ...

Tapping into the power of collaboration, Johns Hopkins University and West Virginia University are launching a new partnership to bring together researchers from both institutions to address complex ...

When Meituan, China’s dominant lifestyle super app that combines services similar to DoorDash, Yelp, and Groupon into a single platform, launched its Xiaomei AI agent in late 2025, executives ...

Available worldwide on Xiaomi’s developer site MiMo Studio, Hugging Face and the API Platform, MiMo-V2-Flash “particularly excels in reasoning, coding and agentic scenarios, while also serving as an ...