AI News
Daily Tech Reader - AI Edition
Apple
Apple
Siri AI EU Launch Delayed Again as Apple Continues DMA Compliance Talks With Brussels
Apple confirmed Siri AI will not launch initially in the EU on iOS, iPadOS, or watchOS, as regulators scrutinize cross-app data access and interoperability requirements under the Digital Markets Act.
Sources: Apple · 9to5Mac · MacRumors
Apple M6 Mac Mini AI Benchmarks Exceed Every Prior Desktop — Starts at $899
Apple's newly released M6 Mac Mini processes large language model prompts 8.5x faster than its predecessor; the M5 Pro configuration is also available, with prices starting $100 higher than the previous generation.
Sources: CNBC · Apple · 9to5Mac
Apple CEO Tim Cook Pledges Disaster Relief Donation for Nepal-Tibet Flash Floods as Company Marks Fifth Crisis Response in Two Months
Cook announced today via X that Apple will donate toward relief efforts after flash floods and mudslides killed hundreds near the Nepal–Tibet border — the company's fifth disaster relief pledge since late June.
Sources: MacRumors · Apple
Microsoft
Microsoft Copilot Mobile Now Sends Proactive Push Notifications — "Your Day at a Glance" Rolls Out Today
The Copilot mobile app begins notifying users proactively for the first time — starting with a daily briefing — marking a shift from a tool you open to one that reaches out on its own.
Sources: A Guide to Cloud · Microsoft
Copilot Canvas Hits General Availability — Word and PowerPoint Get AI-Assisted Tables of Contents, Headers, and Comment Tasks
Microsoft's Canvas document editor is now generally available inside Gemini Enterprise and M365 Copilot, with Word able to auto-build document structure and PowerPoint able to write, reply to, and resolve comments.
Sources: Empowering Cloud · Microsoft
Microsoft Launches "Guild Copilots" — Industry-Specific Copilot Tiers Now Live Beyond Healthcare and Legal Pilots
Following successful sector pilots, Microsoft has formally launched its Guild Copilot tier, bringing domain-tailored AI agents to additional regulated industries as the platform evolves toward ambient enterprise AI.
Sources: AI Conference London · Microsoft
Claude
Anthropic Releases Model Hardware Standard — Claude Can Now Read Robot and Lab Equipment Specs Without Paper Manuals
Anthropic's new MHS protocol lets Claude understand how physical hardware works directly, enabling safer AI integration with robotics, scientific instruments, and manufacturing equipment under a research preview.
Sources: Japan Times · Anthropic
Anthropic Launches Claude Team Plan for Scientists — 10,000 Free and Discounted Seats for Academic and Nonprofit Labs
Starting today, principal investigators at academic and nonprofit institutions can claim free standard seats or $15/month premium seats (80% off) for their research teams at claude.com/programs/team-plan-for-scientists.
Sources: Anthropic · ExplainX · Releasebot
Claude Code Adds Restricted Mode, Cross-Session Messaging, and Enterprise Usage Credits
The latest Claude Code update ships a restricted mode that strips command-execution tools for sandboxed deployments, plus cross-session messaging, keyless Console sign-in, and improved agent workflow and MCP diagnostics.
Sources: Releasebot · Anthropic
ChatGPT
Anthropic Weighs Shareholder Secondary Stock Sales Ahead of Potential $1.5–2 Trillion IPO — Prospectus Expected After Labor Day
Reports indicate Anthropic is in discussions about allowing existing shareholders to sell shares in a secondary offering, as the company's S-1 process advances and bankers discuss a valuation that could set a record for any IPO.
Sources: Tech Startups · Polymarket · NYT
GPT-5.6 Sol API Pricing Slashed More Than 20% for Three Months — Rolls Out Across ChatGPT Work and Codex Plans
OpenAI cut GPT-5.6 Sol API and credit pricing by over 20% effective today for a 90-day window, available on the API and eligible ChatGPT Work and Codex plans; Pro, Plus, and Business subscription usage is unchanged.
Sources: OpenAI · Releasebot
OpenAI Expands in Brazil and Thailand as International AI Push Accelerates
OpenAI announced expanded presence in Brazil and a partnership supporting Thailand's next generation of AI startups today, continuing its international growth push alongside a recent ChatGPT for Academic Researchers program.
Sources: OpenAI · Tech Startups
Gemini
Google Introduces Expert Intelligence — Bring Your Own Google Play Ebooks Into Gemini Notebook
Google's new Expert Intelligence feature lets users pull eligible Google Play ebooks directly into Gemini Notebooks, with third-party subscriptions and broader rollout to AI Mode and the Gemini app coming soon.
Sources: Crypto Integrated · Google
Gemini 3.6 Flash and 3.5 Flash-Lite Hit General Availability — Robotics Model Deprecated August 31
Google moved Gemini 3.6 Flash (improved token efficiency, better agentic planning) and 3.5 Flash-Lite (ultra-low-latency for high-volume automation) to stable GA today; the gemini-robotics-er-1.6-preview shuts down Sunday.
Sources: Google AI for Developers · Releasebot
OpenAI, Anthropic, Google, Microsoft, and 100+ Companies Sign Joint Warning on AI-Powered Cyberattacks Against Critical Infrastructure
The coalition warns that hospitals, water systems, and internet infrastructure face growing risk as AI models improve at discovering vulnerabilities and autonomously executing attacks — calling for government and industry coordination on new defenses.
Sources: TechCrunch · Tech Startups
China AI
DeepSeek Releases Harness — Its First Agent Runtime Framework for Direct Code Repository Interaction
DeepSeek's new Harness framework lets AI models locate files, modify code, run tests, and fix errors iteratively inside code repositories — a direct move into agentic software development tooling.
Sources: Global Times · DeepSeek
Alibaba's Qwen3.8-Max Reaches 2.4 Trillion Parameters — Largest Chinese AI Model to Date Now Competing With OpenAI and Anthropic Frontiers
Alibaba's Qwen3.8-Max uses a mixture-of-experts architecture activating ~95 billion parameters at a time, supporting a 1-million-token context window and claiming benchmark parity with leading U.S. frontier models.
Sources: Bloomberg · The Register · CTech
Big Story
Anthropic Wins Federal Court Battle Against Pentagon — Judge Rules Defense Department Unlawfully Retaliated After Claude's Military AI Limits Were Enforced
A California federal judge ruled the DoD violated the law by designating Anthropic a supply-chain risk after the company refused Pentagon contracts that would have overridden Claude's red lines on lethal autonomous weapons and domestic mass surveillance.
Sources: Financial Times · Tech Startups
a16z Launches $1.1 Billion Machine Age Fund — AI Investment Shifts From Software to Chips, Robots, and Physical Infrastructure
Andreessen Horowitz's first hardware-dedicated fund targets processors, memory, networking, storage, and robotics, signaling that AI's biggest remaining bottlenecks — and its next major fortunes — now live below the application layer.
Sources: Wall Street Journal · Tech Startups