The news desk
Recent developments in artificial intelligence, with the original sources one click away.
Apple's iOS 27.3 developer preview lets Siri AI perform actions in third-party apps using natural language and on-screen context. Developers can create combined-task shortcuts through a new app skills feature.
Sources · theverge.com
The cybersecurity-focused VC firm filed to raise $360M for its third fund and $130M for a separate vehicle, TechCrunch reports, citing SEC filings.
Sources · techcrunch.com
Google Antigravity now includes cheminformatics skills that integrate RDKit, ChEMBL and PubChem to automate molecular analysis tasks. DeepMind reports the agent completed a workflow involving matched molecular pair analysis in about an hour.
Sources · deepmind.google
Google will entirely disable Search access for K-12 student accounts if administrators turn off Gemini AI features, the company told The New York Times, saying the services are now "deeply integrated."
Sources · theverge.com
Google plans to make its AI tools opt-in for K-12 schools in 2027, but schools that disable AI will lose access to Google Search because AI Overviews are now integrated into the search engine.
Sources · techcrunch.com
Anthropic modified Claude Code's system prompt to steer the AI away from recreating specific existing games, while still allowing users to build original games and genre-based titles.
Sources · theverge.com
Google released Gemini 4 Argon globally in October for coding and analysis tasks, then announced Gemini 4 Carbon for cybersecurity. The company also introduced Genie 4 for real-time world simulation with audio.
Sources · blog.google
NVIDIA announced partnerships with cities and developers in Spain, Poland, Malaysia and the San Francisco Bay Area to deploy AI infrastructure for resident-service agents and camera-feed analysis using its Nemotron, Cosmos and Metropolis tools.
Sources · blogs.nvidia.com
OpenAI's chief global affairs officer told Axios the company is planning communications efforts around the midterm elections, including proposals to rebrand data centers and develop AI regulation.
Sources · techcrunch.com
OpenAI is combining ChatGPT's Chat and Work tabs in the desktop app, initially for select users. The change consolidates project workflows while keeping Codex separate, but will pool token allowances across both interfaces.
Sources · techcrunch.com
Mount Shasta wilderness managers report multiple incidents of hikers relying on AI chatbots that provided incorrect trail information, camping guidance and maps, though annual rescue totals remain steady.
Sources · wired.com
A federal court rejected xAI's motion to dismiss an NAACP lawsuit alleging the company operated gas turbines at its Memphis-area Colossus data center without required air pollution permits.
Sources · techcrunch.com
Google reports its experimental satellite equipped with Tensor Processing Units successfully ran the Gemma 4 model in space, generating a haiku and completing other prompts as part of tests for potential space-based AI data centers.
Sources · theverge.com
Researchers found new malware called Volta-Autodoxer that uses commercial LLMs to analyze stolen credentials and prioritize high-value targets, according to findings shared with Wired.
Sources · wired.com
OpenAI has reduced its fundraising goal from a previously reported $30 billion to $20 billion, with $11 billion secured so far, according to reports from the Financial Times and Wall Street Journal.
Sources · techcrunch.com
An AI agent controlled a robotic surgical rig to inject gene therapy into a mouse's brain after observing a veterinarian's initial procedure, demonstrating direct AI participation in laboratory work.
Sources · wired.com
A malicious version of Anthropic's official Claude extension was distributed through Cursor's marketplace, installing SSH backdoors on developer machines. Anthropic has not issued a public warning.
Sources · arstechnica.com
OpenAI's ChatGPT desktop app, including the Codex coding agent, can now route inference through Amazon Bedrock in AWS GovCloud (US), allowing organizations to meet data residency and security requirements with IAM controls and zero data retention.
Sources · aws.amazon.com
Four major AI companies say they will adopt a proposed shared signal in robots.txt files that lets publishers set AI training preferences once across multiple providers, instead of configuring each company separately.
Sources · theverge.com
The SETI Institute and Breakthrough Listen are using AI agents to steer the Allen Telescope Array, process incoming data and identify potential extraterrestrial signals from natural language requests.
Sources · blogs.nvidia.com
