👀In Case You Missed It
OpenAI and Jony Ive are developing a new "family of AI devices" focused on calmer, healthier human-AI interaction, but technical challenges mean the launch will take time.
Zendesk’s new AI agents, including voice assistants and copilots, claim up to 80% support ticket resolution, showing strong AI adoption in customer service.
Apple warns of privacy risks as it prepares to comply with Texas’s age-assurance law, highlighting regulatory challenges.
India pilots AI chatbot-led e-commerce using ChatGPT, Gemini, and Claude models, signaling AI integration in emerging market retail.
Microsoft partners with Harvard Medical School to license medical content, improving reliability of Copilot’s health answers.
Read more here
🧠 Cloud & Infrastructure Watch
Meta talent coup: Thinking Machines Lab co-founder Andrew Tulloch exits to Meta’s Superintelligence Lab amid a broader AI team reorg and huge infra spend.
Read MorexAI world-model push: Hiring ex-Nvidia researchers to build interactive, AI-generated 3D game worlds; Musk teases a playable title by 2026.
Read MoreGPT-5 bias audit: OpenAI says GPT-5 “instant/thinking” cut political bias ~30% across 500 contentious prompts; strongest gains on emotionally loaded queries.
Read MoreMicrosoft capacity squeeze extends into 2026 — internal forecasts show Azure shortages in key US regions (NoVA, Texas), restricting new subs and rerouting workloads despite massive build-outs.
Read MoreSalesforce Agentforce 360: New push to let enterprises build/deploy task-doing agents across CRM workflows.
Google leak points to Gemini 3 on Oct 22 with stronger multimodal reasoning.
💻 Hardware & Semiconductor Moves
Intel 18A chips: New processors fabbed on 18A process (Arizona) as Intel leans into US manufacturing.
Read MoreAnduril “EagleEye” MR helmet: Soldier-worn, AI-augmented mixed-reality system.
Read MoreAI edge optics: “AI cameras” arms race for real-time inference; new hardware vuln research shows training data can be hacked via silicon flaws.
Read More
