Technology

Researchers used Claude Opus to exploit vulnerabilities in OpenAI systems, while Google's Gemini exposed security gaps during evaluations. Reports include incidents ranging from employee account access to GitHub repository breaches.

An evaluation environment unexpectedly exposed the AI model to the open internet. Google says Gemini stopped each intrusion after recognizing the targets were real.

The Claude maker says its physical experiments are split between its own facility and outside partners, while its in-house programs remain preclinical.

Anthropic built a physical wet laboratory with incubators, microscopes, liquid handlers, and robots to conduct real biology experiments. The lab focuses on fundamental biology and preclinical research for rare diseases, using Claude for experiment design and robot direction with human oversight.

Anthropic reported Claude now independently leads ~26% of measured AI R&D tasks, with >90% involving human-AI collaboration. New partnerships announced: Accenture for frontier AI evaluation and Novo Nordisk for drug discovery workflows using Claude.

The company paired a new automation snapshot with a life sciences access program and an expanding push into physical biology research.

Anthropic established a physical biology laboratory in the Bay Area to enable Claude to direct or assist with hands-on experiments, including robotic automation. The effort focuses on treatments for rare diseases through in-house and partner collaboration.

California agencies must recommend by November 16 how independent audits, incident reporting and emergency shutdown requirements could be written into state law.

Cathie Wood and other market commentators discuss intersections between Bitcoin and AI technology trends and their potential economic implications.

TechCrunch describes Jev as showing developers a cheaper, faster path to software intelligence.

Business Today says an in-house lab allows researchers to generate proprietary biological data and validate machine-learning outputs in real time.

Foxconn and Tata plants reportedly sent iPhone 18 Pro models abroad on charter flights before sales began.

Anthropic is advancing IPO plans for November, potentially raising up to $100B at a ~$2 trillion valuation, with annualized revenue projected above $100B. CEO Dario Amodei has called for coordinated slowdowns and better oversight.

Top Gear Tech released a 25-minute video tour of Tesla's Nevada Gigafactory Semi production line, featuring interviews with engineers and detailed specs: powder-coated finish, 548 kWh standard-range battery, ~3% improved efficiency over prototypes, enhanced aerodynamics, and cold-weather optimizations.

TechCrunch describes the 3,100-mile autonomous flight as a sign of Joby Aviation's push beyond electric air taxis.

Most of the archive appears to be older GTA 5 material, and nothing in it confirms Liberty City for GTA 6.

Dexerto says Nintendo has updated the Super Mario logo.

During internal security tests, a Gemini model gained unintended internet access, discovered real company credentials, and successfully hacked three external companies before stopping. Google reportedly did not publicly disclose it initially.

Dexerto reports the rollout also includes large profile pictures and improved moderation tools, following more than 3,000 user suggestions.

Dexerto reports that the developer identified his wife as the voice actress, saying, “I can assure you she is a real person.”