AI News
Daily Tech Reader - AI Edition
Podcast 🎧 • Video 📽 • Short 📽
Apple
Apple
Apple Seeds iOS 27 Beta 7 to Developers — Siri AI Refinements and Performance Gains Ahead of September Launch
Apple released the seventh developer beta of iOS 27, iPadOS 27, macOS 27, and companion OS releases today, focusing on stability, Siri AI conversational refinements, and Dynamic Island integration as the final public release targets September 14.
Sources: MacRumors · 9to5Mac · AppleInsider · MacObserver
Gemini 3.7 Flash Now Powers Apple's AI Mode Search — Siri AI Expansion Still Blocked in EU
Google's Gemini 3.7 Flash is now driving Apple Intelligence features in iOS 27 beta, while Siri AI remains blocked for EU users on iPhone and iPad due to ongoing DMA compliance delays with no resolution date confirmed.
Sources: 9to5Mac · FelloAI · Apple
Tim Cook Celebrated at Apple as He Prepares to Step Down September 1 After 15 Years as CEO
Apple held an internal celebration Saturday for outgoing CEO Tim Cook, attended by roughly 200 people, ahead of his formal departure September 1 — closing a 15-year tenure that encompassed the rise of Apple Intelligence.
Sources: AppleInsider · Bloomberg
Microsoft
Microsoft 365 Copilot Hits 30 Million Paid Seats — OpenAI Quietly Becomes a Subprocessor in Tenant Data
Microsoft closed FY26 with 30 million paid Copilot seats and Azure up 43%, while quietly auto-enabling OpenAI as a subprocessor for GPT-5.6-class models in Microsoft 365 — a toggle administrators need to review immediately.
Sources: Empowering.cloud · Microsoft
Dragon Copilot Physician Apps Launch in Microsoft Marketplace — Partners Can Embed AI Into Clinical Workflows
Microsoft opened Dragon Copilot to healthcare partners via a new Marketplace program, letting software companies embed specialty AI capabilities directly into physician workflows without custom integration work.
Sources: Microsoft Learn · Microsoft Partner Center
Microsoft 365 Copilot August Update: GPT-5.6 and Claude Models, Cowork Integration, Expanded Notebooks
The August M365 Copilot release adds dynamic AI model routing between GPT-5.6 and Claude, a new Cowork integration for custom workflow automation, and expanded Copilot Notebooks access for all users.
Sources: Geeky Gadgets · Microsoft · Releasebot
Claude
Claude Outage Hits Multiple Models at 5 AM UTC — Mythos 5, Fable 5, Opus 5, and Opus 4.8 All Affected
Anthropic's status page flagged elevated errors across Claude Mythos 5, Fable 5, Opus 5, and Opus 4.8 starting at 05:06 UTC today, impacting claude.ai, the API, Claude Code, and Cowork — with remediation ongoing past 06:42 UTC.
Sources: Anthropic Status · Arabian Business · CybersecurityNews · DesignTAXI
Anthropic Public S-1 IPO Filing Expected as Soon as This Week — October Debut Targets $2 Trillion Valuation
Sources tell Bloomberg and CryptoBriefing that Anthropic is preparing to publicly file its S-1 registration as early as this week, following a confidential June 1 filing, with investor discussions pointing to an October listing at a $2 trillion-plus valuation.
Sources: Bloomberg · CryptoBriefing · SmartAsset · UnusualWhales
Anthropic Introduces AI Watermarking for Claude — EU Transparency Requirements Trigger August Rollout
Anthropic is rolling out machine-readable watermarks for Claude-generated content to meet EU AI Act transparency mandates, with new Claude models in EU markets required to support the marking starting August 2.
Sources: The Social Butterfly · Anthropic
ChatGPT
ChatGPT Ads Go Live Across 31 European Markets Today — Cost-Per-Click Now Dominates Ad Spending
OpenAI launched ChatGPT advertising across 31 European markets on August 24, extending a US ad business that launched six months ago, with European advertisers initially buying through major agency groups and self-serve access planned later.
Sources: MarketingProfs · OpenAI · TechCrunch
OpenAI o3 Retires From ChatGPT on August 26 — GPT-5.4 mini Now Available to Free Users via Thinking Feature
OpenAI's o3 model exits ChatGPT in two days following a 90-day sunset, while GPT-5.4 mini rolls out to free and Go users through the Thinking feature as OpenAI continues consolidating its model lineup around the GPT-5.x family.
Sources: OpenAI · Releasebot
ChatGPT Desktop App Launches on Linux — Now Available on Ubuntu, Debian, and Fedora in Public Preview
OpenAI released the ChatGPT desktop app for Linux in public preview today, supporting Ubuntu 24.04 and 26.04 LTS, Debian 13, and Fedora 43/44, with browser actions available but desktop app control not yet supported.
Sources: OpenAI · Releasebot
Gemini
Gemini 3.7 Flash Now Powers Gemini Spark Across 160+ Countries — Updated CBRN and Cyber Safety Guardrails Ship
Google upgraded Gemini Spark to run on 3.7 Flash today, improving accuracy for complex multi-skill Workspace workflows, and shipped updated safeguards against CBRN and cyber-offense misuse alongside the model rollout.
Sources: Google · Releasebot
Gemini Hits 1 Billion Monthly Users as Google Workspace Gets Admin Assist and New DLP Controls
Google confirmed Gemini surpassed 1 billion monthly active users in August, while rolling out Admin Assist with a Gemini-powered Admin Console sidepanel and new data loss prevention controls for Drive and Studio flows.
Sources: Google · Science Online · Google Workspace Updates
Ask Gemini Launches in Google Chat on August 26 — Replaces Side Panel With Unified Workspace Command
Google's Ask Gemini rolls out to Google Chat in two days, replacing the existing Workspace side panel with a single command for search, drafting, task management, and conversation catch-up using full Workspace context.
Sources: Google · Releasebot
China AI
Chinese Hackers Are Weaponizing DeepSeek and Open-Source AI to Escalate Cyberattacks Against Western Targets
Bloomberg reports today that Chinese state-linked hackers have integrated DeepSeek and other open-source AI models into offensive operations, using the tools to accelerate attack planning and targeting against organizations abroad.
Sources: Bloomberg
Hugging Face Exploring $13 Billion Sale as AI Platform M&A Heats Up
Hugging Face, the open-source AI platform and model hub, is reportedly exploring a sale at a valuation around $13 billion, according to sources cited today — a deal that would mark one of the largest AI platform acquisitions to date.
Sources: Crypto Integrated · AI Weekly
Big Story
Anthropic's S-1 Week: Why the IPO Filing Lands in the Middle of a Reliability Crisis
The same week Anthropic is expected to publicly file its S-1 targeting a $2 trillion valuation, Claude suffered its eighth significant outage of August — a tension between the explosive growth narrative the IPO requires and the operational maturity public investors will scrutinize.
Sources: Bloomberg · CybersecurityNews · CNBC · CryptoBriefing
Open-Source AI as a Weapon: DeepSeek Report Reframes the China AI Threat From Competition to Offense
Bloomberg's report that Chinese hackers are actively weaponizing DeepSeek shifts the AI-China narrative from a race about model performance to a live national security threat — with open-source access now framed as an attack surface, not just a competitive advantage.
Sources: Bloomberg · AI Weekly · CNBC