xAI develops Grok-powered advanced file editor with spreadsheet capabilities, integrating AI into productivity tools to challenge OpenAI, Google, and Microsoft. Launched April 2025 Grok Studio for collaborative document/code generation, competing with Gemini workspace.

Google enhances Chromebook Plus with AI features including screen capture/search tools, text simplification, and Gemini integration. New models offer AI image creation, 2TB storage, and $649-$749 OLED displays with MediaTek chips, targeting productivity markets.

Tesla's autonomous taxi pilot in Austin, using 10 Model Y vehicles with FSD tech, drives over 8% stock surge amid NHTSA investigations into safety concerns. The company's camera-based system contrasts with competitors' multi-sensor approaches. Expansion plans aim for 1,000 vehicles by year-end before LA/SF rollout.

OpenAI removes io-branded content after trademark dispute with iyO, impacting hardware plans led by Jony Ive and facing legal challenges, partnership shifts, and Meta talent competition.

Microsoft unveils Mu, a compact on-device language model for Windows 11 Settings enabling natural language system configuration. The encoder-decoder architecture achieves >100 tokens/second on NPUs through parameter alignment, memory optimization, and quantization techniques. Currently available in Windows Insider Preview builds for Snapdragon X-series Copilot+ devices, with AMD/Intel compatibility planned for future updates.

X Corp accelerates X Money's physical card development, aiming U.S. launch in 2025 with global expansion in 2026, supported by 40+ state licenses and potential partnerships with Visa, Mastercard, and Amex.

LinkedIn CEO Ryan Roslansky noted AI content polishing faces unexpected adoption challenges on professional networks due to users' heightened sensitivity to AI-generated traces impacting career credibility, despite 600% growth in AI job demands and 20x increase in skill updates, as professional content equals a digital resume where AI markers could jeopardize real-world opportunities.

Anthropic enhances Claude Code with remote MCP server connectivity, streamlining external tool integration via OAuth authentication (198 chars).

Google launches Android 16 with new APIs, adaptive UI for large devices, enhanced notifications, and security upgrades. Developers must update apps for compatibility. Material 3 Expressive UI follows later. Key features include desktop-style multitasking, improved privacy controls, and USB attack protection. (199 chars)

Meta and Oakley unveil Oakley HSTN smart glasses with enhanced AI, 2880x1620 camera, 8-hour battery, and IPX4 water resistance, priced at $399 for six frame styles. Research-oriented Aria 2 features depth-sensing cameras and offline AI processing for developers.

MIT study shows LLM-assisted writing reduces cognitive engagement and neural connectivity. AI users showed lower theta wave activity and citation accuracy, raising concerns about learning capacity degradation.

Purdue researchers develop a training-free robot inspection method using pre-trained VLMs to convert natural language into optimized 3D trajectories for infrastructure monitoring, achieving over 90% spatial relationship accuracy through TSP optimization and semantic alignment.

Gemini now enables direct video upload and analysis on Android/iOS, offering contextual insights from personal videos. Web support and in-app recording are pending, but it outperforms rivals with precise object/action recognition and temporal analysis.

Microsoft unveils Microsoft 365 Agent Toolkit, enabling developers to build AI-driven workflow automation across M365 platforms. The toolkit offers declarative agent creation, multi-format deployment (VS Code, VS, GitHub Copilot, CLI), and SDKs for .NET/Python/TypeScript integration, streamlining intelligent agent development and deployment via Microsoft Graph and Azure services.

Meta expands AI team via high-profile hires like ex-GitHub CEO Nat Friedman and Daniel Gross, acquires Scale AI for $1.43B to build AGI capabilities, intensifying competition with OpenAI/Google through $100M+ sign-on bonuses in talent war.

OpenAI uncovers AI model features tied to harmful behaviors, enabling manipulation to mitigate risks. Safety training realigns models, advancing safety research through improved interpretability.

Google launches Gemini 2.5 AI models for production use, featuring Flash-Lite as the fastest/cost-effective variant in preview mode, all built on Mixture-of-Experts architecture for optimized efficiency in large-scale applications.

Google introduces Live Search, enabling dynamic voice conversations in AI Mode for real-time, multi-step queries. The feature uses a custom Gemini model with web-linked responses, rivaling ChatGPT and Perplexity AI. Users access it via the Google app, with future plans for camera-based context.

OpenAI terminates Scale AI partnership, shifts to new data providers amid Meta's investment and leadership changes.

Midjourney unveils AI video model V1 for image-to-video generation via Discord, competing with Sora/Gen4. Legal challenges from Disney/Universal follow copyright claims. Video generation costs 8x image rates, with $10-$120 monthly plans. The startup aims to develop 3D simulations while facing Hollywood's AI regulation concerns.