ChatGPT and Google AI tools increasingly cite Grokipedia, raising concerns over misinformation as its usage grows despite being a secondary source compared to Wikipedia.

Anthropic enhances Claude Cowork with custom plugins, enabling automation of tasks like CRM data extraction and sub-agent creation. NASA used Claude to halve Mars rover route planning time.

SpaceX seeks FCC approval to launch a million orbital data center satellites, proposing a laser-linked, solar-powered network as a more sustainable alternative to Earth-based AI infrastructure.

DeepSeek OCR 2 introduces a vision encoder that reorganizes image information semantically, not sequentially. It replaces a standard CLIP module with a compact language model, using fewer tokens (256-1120 vs. 6000+) and improving performance on benchmarks. This marks a step towards unified multimodal processing.

RobCo secures $100 million Series C funding to advance Physical AI robotics for U.S. and European manufacturing automation.

Google is developing a voice cloning feature in AI Studio, linked to Gemini models. A "Create Your Voice" option has been spotted for recording audio, though non-functional, hinting at future integration for Gemini 3 Flash to generate synthetic voices. The December 2025 update improved Gemini 2.5 Flash's audio quality and instruction following. New options for importing GitHub code repos and a revised start page with separate activity stats are also in the works.

Apple's second-largest acquisition is AI audio startup Q.ai, reportedly for $2 billion. Its optical sensor tech could enable non-verbal AI assistant communication via headphones or glasses, potentially enhancing Siri and Apple devices like AirPods and Vision Pro.

Google Maps now offers hands-free Gemini AI for walking and cycling, enabling real-time conversational navigation and queries. The update is rolling out globally on iOS and soon on Android.

Music publishers sue Anthropic for illegally downloading over 20,000 copyrighted songs, seeking over $3 billion in damages. This follows a prior $1.5 billion penalty for similar copyright infringement in a separate author lawsuit.

Google DeepMind opens access to Project Genie, an AI tool that creates interactive game worlds from text or images. Available to U.S. Google AI Ultra subscribers, it uses the Genie 3 world model and aims to gather feedback for developing more powerful models. While seen as a step toward AGI, initial applications focus on entertainment, with plans to expand to training robots in simulations. The tool excels in artistic styles but struggles with realism and precise navigation, reflecting its exper

Lorikeet launches Coach, an AI analytics assistant for regulated industries that diagnoses AI support issues via topic analysis and ticket quality scoring, providing full coverage and actionable insights beyond traditional CSAT metrics.

Elon Musk's SpaceX, xAI, and Tesla are in early merger talks. Options include SpaceX merging with either Tesla or xAI, potentially consolidating before SpaceX's IPO.

Alibaba's T-Head launched the fully self-developed "Zhenwu 810E" AI chip, completing its AI "golden triangle" with Tongyi Lab and Alibaba Cloud. Its performance rivals NVIDIA's H20, surpassing the A800 and domestic GPUs.

OpenAI and Anthropic are seeking new funding rounds, potentially their last before IPOs. OpenAI could raise up to $30B from SoftBank, valuing it at $830B. Anthropic is raising $20B, valuing it at $350B. Both target IPOs in 2026.

Google launches "Auto Browse" in Chrome, powered by Gemini AI, to automate multi-step tasks like bookings, form filling, and shopping. It integrates with Gmail, Calendar, and more, moving to a side panel. The update also includes Nano Banana for image generation and plans future personal intelligence features.

Google DeepMind's Gemini 3 Flash gains "Agentic Vision," enabling iterative image analysis via code execution for tasks like zooming and annotation, improving accuracy. Features require prompts; rollout ongoing.

AI startup Decart launches Lucy 2.0, a real-time video transformation model. It enables live 1080p video editing at 30fps with near-zero latency, allowing character swaps, product placement, and environment changes via text and image prompts. The model runs on AWS Trainium3 chips.

Meta Q4 results beat forecasts with $8.88 EPS and $59.89B revenue, up 24% YoY. Q1 revenue guidance of $53.5B-$56.5B exceeds expectations. AI and metaverse investments drive higher 2026 capex of $115B-$135B. First AI model from Superintelligence Labs expected soon.

Meta to charge developers for AI chatbot services on WhatsApp Business API in Italy from February 16, following regulatory pressure. Fees set at $0.0691 per AI-generated response, with potential expansion to other regions.

Tesla invested $2B in xAI despite shareholder rejection, citing strategic alignment with its AI-focused Master Plan Part 4 for physical-world AI and robotics collaboration.