Tech News

Story of the Day

OpenAI announced it has solved the 90-year-old Navier-Stokes Millennium Prize problem using an unreleased internal AI model alongside 10,000 concurrent agents. The claim sparked immediate controversy and sharp pushback from academic mathematicians, who questioned the proof’s validity, methodology, and lack of peer verification.

Top Stories

Apple Set to Launch First Foldable iPhone Under CEO John Ternus · The Verge Apple is expected to unveil its long-rumored foldable iPhone at its annual launch event, marking the company’s first major hardware form-factor shift in years and the debut flagship launch under new CEO John Ternus. The high-end device, which has been in development for nearly a decade and is rumored to carry a price tag exceeding $2,000, enters a market where competitors like Xiaomi and Samsung are already shipping squat, wide-screen foldables. While the premium Pro and Fold models are slated to ship first, standard iPhone 18 base models have reportedly been delayed until early next year.

Meta Launches Autonomous Personal AI Agent Muse · The Verge Meta introduced Muse, a personal AI assistant operating inside a dedicated cloud virtual machine equipped with a built-in browser to perform real-world tasks autonomously. Directed by simple prompts, Muse can execute multi-step workflows like booking travel, negotiating purchases, and sending emails by interacting with websites and filling out forms directly. To address privacy concerns surrounding autonomous execution, Meta equipped Muse with granular user controls, including a planned confidential mode that prevents activity from being viewed by Meta itself.

Google DeepMind Maps All 9 Billion Human Genome Mutations · Slashdot Google DeepMind released AlphaGenome Atlas, a comprehensive database predicting the biological effects of all 9 billion potential single-letter DNA mutations across the human genome. The tool evaluates how genetic variations alter gene expression and regulatory machinery, replacing labor-intensive lab testing with instant browser-based lookups for researchers. DeepMind made the database freely accessible for non-commercial research, with commercial licensing planned via Google Cloud.

US Intelligence Accuses Chinese Firms of Systematic AI Distillation · Slashdot A joint advisory from the NSA, CISA, and FBI accuses Chinese AI developers of conducting industrial-scale knowledge distillation campaigns against leading American AI models. According to intelligence officials, companies like DeepSeek and Moonshot AI systematically routed millions of queries through proxies, cloud providers, and grey markets to extract capabilities from ChatGPT, Claude, Gemini, and Grok. The harvested outputs were allegedly used as the primary training foundation for domestic Chinese models like DeepSeek-R3 and Moonshot’s Kimi-K3.

Qualcomm Inks Multi-Generation Custom AI Chip Deal with Amazon · Bloomberg Qualcomm signed a multi-generation partnership to supply custom data center silicon for Amazon Web Services’ expanding AI infrastructure. As part of the agreement, Amazon receives warrant rights to purchase up to $4 billion in Qualcomm shares as deployment milestones are met. The strategic deal allows Amazon to diversify its hardware supply away from Nvidia while providing Qualcomm a major entry point into server-grade AI hardware.

AI-Driven Bug Hunting Triggers Record 972 Microsoft Vulnerability Patches · Ars Technica Microsoft issued a massive Patch Tuesday security update addressing a record 972 vulnerabilities, including 112 rated at critical severity. Security engineers attributed the unprecedented surge in reported flaws to advanced AI models rapidly uncovering vulnerabilities across operating systems and browsers. The fix comes shortly after major tech and cybersecurity firms issued a joint warning regarding a narrowing timeline to patch software before AI-driven exploits proliferate.

Mistral AI Secures €3 Billion Funding Round at €21 Billion Valuation · TechCrunch French AI lab Mistral raised €3 billion in Series D funding led by Samsung Electronics, pushing its valuation to €21 billion in the largest private tech round in European history. The company plans to deploy the capital into building compute infrastructure and advancing open-weights AI models. The raise reflects growing European demand for sovereign AI capabilities alongside a broader strategic pivot by Mistral toward data center infrastructure.

AI Coding Pioneer Cognition Hits $48 Billion Valuation · TechCrunch AI software engineering startup Cognition raised $2 billion in new capital, vaulting its valuation to $48 billion. The massive valuation multiple highlights fierce investor conviction that the automated coding sector remains open to high-growth challengers rather than consolidating into a single winner. Cognition’s surge comes as software engineering teams increasingly adopt agentic workflows to accelerate code production and maintenance.

Also Worth Knowing

• LG Smart TVs Recorded Audio and Scanned Local Networks in Standby Mode (The Verge): Independent security investigations revealed LG OLED TVs continuously scan local Wi-Fi networks for nearby devices and log microphone audio even when offline or placed on standby mode.
• Australia Proposes Bill Allowing Users to Disable Algorithmic Social Feeds (Slashdot): Proposed Australian legislation would force social media platforms to let users over 16 opt out of algorithmic recommendation feeds and return to chronological posts from followed creators.
• US Energy Department Approves $1.9 Billion Loan to Restart Nuclear Reactor for Google Data Centers (Engadget): Federal regulators approved a $1.9 billion loan to revive a shuttered Iowa nuclear plant to supply dedicated zero-emission power for Google AI data centers.
• Google Degrades European Travel Search Results to Comply with EU DMA Rules (Ars Technica): Following a €460 million antitrust fine, Google altered European search layouts to prioritize third-party travel aggregators over direct local business links while warning the change degrades user experience.
• OECD Flagship Study Links Student AI Chatbot Use to Lower Academic Test Scores (Bloomberg): An international PISA educational assessment found 15-year-old students who regularly use AI chatbots for schoolwork scored roughly 20 points lower in science than non-users.

💡 Next Step: Would you like me to generate a audio overview or a slide deck summarizing today’s top stories for a quick team briefing?