Technology

Google says Extended Thinking can talk through longer tasks as it works, with rollout beginning in Gemini Live, developer tools and selected Workspace features.

Core and Premium combine paid features across Instagram, Facebook and WhatsApp. Meta says the apps’ core experience will remain free.

The company says it banned accounts, but has not established harmful intent or an imminent biological threat.

Meta is advancing custom MTIA silicon, planning to deploy MTIA 450 (Arke) in H1 2027 and MTIA 500 (Astrid) by end-2027. These target efficient inference workloads with better performance-per-watt and cost than Nvidia systems. Meta partnered with Broadcom and TSMC.

The September 14 updates address Remote Desktop failures, Linux shared folders and multichannel audio. Microsoft says some USB audio devices still may not work.

Apple rolled out iOS 27 with a major Siri overhaul powered by Apple Intelligence, offering conversational and context-aware responses. New child safety features include age-based app controls, Ask to Browse for website approval, Communication Safety detection, and Time Allowances for daily usage limits.

Apple released iOS 27 with conversational Siri AI featuring personal context awareness, app actions, and onscreen understanding. Supports iPhone 11 and later, with advanced AI on iPhone 15 Pro/16+ models, plus performance gains and Liquid Glass UI refinements.

Meta introduced Muse, a proactive personal AI agent handling email, travel booking, payments, and workflow automation across connected apps. Running in a secure VM with privacy controls, it is available in the US via app, WhatsApp, and web with free and subscription tiers ($20/$100).

The Hill reports a 49–50 vote on a motion to proceed, with all Democrats and a few Republicans blocking the cryptocurrency regulation bill from advancing.

Apple’s rebuilt assistant starts in English as a beta, requires a newer iPhone and initially excludes iPhones and iPads in the EU.

Google DeepMind announced Gemini 3.8 Live, a native speech-to-speech model with low-latency voice interactions, and Extended Thinking for complex multi-step reasoning. Both support visual context, multilingual switching, and agentic workflows available via Gemini API and Google AI Studio.

Pulse 2.0 reports that the cybersecurity company is expanding its platform for protecting robots, drones, autonomous vehicles and other intelligent machines.

The account-free X reader cites a new legal development. Nitter’s repository is read-only, but permanent closure remains unconfirmed.

The redesigned interface debuts in the 2027 Chevrolet Silverado and GMC Sierra, Ars Technica reports. Engadget describes split-screen support for CarPlay and Android Auto.

SecurityWeek reports 200 vulnerabilities patched with the iOS 27 and macOS Golden Gate 27 releases. iPhone in Canada says Canadian users must join an official waitlist to activate Siri AI.

Microsoft revealed a San Francisco event on how local AI will shape the PC, featuring CEO Satya Nadella, Windows chief Pavan Davuluri, and Nvidia CEO Jensen Huang. Expectations include Windows improvements, new Surface devices with NPU capabilities for local AI, and Arm/Windows progress.

Latent Space previews a conversation with Richard Socher about the lab’s plans, problems with AI peer review and how much remains before artificial general intelligence.

GitLab has released fixes for CVE-2026-85706. Administrators running their own servers need to update; GitLab-hosted services are already patched, the company says.

Business Insider reports that Agility shared renderings of the teal-blue robot, which is supposed to detect nearby people and slow down or stop when they get too close.

Tom's Hardware reports that natural gas consumption by US AI data centers is projected to grow by 15 billion cubic feet per day as demand for computing increases.