The changes that matter, selected from primary sources and summarized with the vendor claim separated from our context.
OpenAI outlines advertising products, GitHub broadens AI Scan eligibility, Google introduces dialogue models, and Jensen Huang argues for vendor-led safety.
The available sources cannot support three substantive stories within the publisher limit. GitHub provides two concrete product updates; DeepSeek offers too little detail, and the remaining entries are customer stories.本
GitHub expands agent workflows, OpenAI reports infrastructure scale, and Anthropic establishes a US security advisory council.
Protocol governance, coding infrastructure and new ChatGPT workflows lead the day, alongside a nuclear-safety classifier whose evaluation details need scrutiny.
Xcode integrates the Claude Agent SDK, GitHub adds centrally managed agent permissions, and Anthropic reports a $30 billion funding round.
GitHub expands JetBrains controls and context, ByteDance introduces EdgeBench, and Anthropic describes an AI-orchestrated espionage campaign.
DeepMind maps predicted DNA-variant effects, Mistral reports a €3 billion funding round, and Anthropic examines security frameworks against AI-enabled threats.
Anthropic plans an independent review, ByteDance releases Seedance 2.5, and OpenAI shares early evidence on coding agents in research.
Anthropic reports unauthorized access during cybersecurity evaluations, GitHub adds GPT-6 Astra, and Claude prepares text watermarking.
Four primary-source updates covering a frontier-model safeguard threshold, Copilot model changes and a new Chinese multimodal model.
Four primary-source updates on cyber defense, weather forecasting, agentic workflows and browser-side AI.