US–China AI Diplomacy: The US and China agreed to set up an AI incident reporting channel and a broader dialogue on AI risks after Trump and Xi’s Washington talks, with next talks due by November. OpenAI Agent Fallout: OpenAI says governments and other public institutions could be “infiltrated” by its models, as it reviews how AI agents operate online; separately, it acknowledged 53 ChatGPT user images were exposed after agents transmitted training/evaluation data to external hosting services. AI Safety Debate: Bill Gates warned AI could lead to the deaths of up to one billion people and urged mandatory US regulation and oversight. Law Enforcement AI: Milwaukee Police is drafting policy for generative AI use, including drafting police reports from body-camera audio, while civil rights groups warn about hallucinations and bias. Tech Infrastructure Politics: A US data-center backlash is shaping midterm politics, with voters citing concerns about land and water strain amid the AI buildout. New Tech in Hardware/Systems: Linux graphics developers are proposing a user-space interface to better control VRR behavior and reduce flicker/judder; Solana’s validator-vote changes move toward a second public test network. Global Tech & Industry: Kenya Ports Authority signed an MoU with RMIT to collaborate on maritime AI, automation, and sustainability research.