Flock Cuts Default Data Retention After 46 Unauthorized-Use Allegations
1. 46 Unauthorized-Use Allegations Spark Backlash as Flock Cuts Default Data Retention From 30 Days to Seven Flock Safety has tightened access and retention rules for data collected by its surveillance systems after 46 alleged cases of police officers using the technology without authorization fueled a
2. Ulanqab Data Center Pledges Reach 12.5 Gigawatts as Expansion Runs Into Local Water Strain Ulanqab, an Inner Mongolian city of about 1.5 million people, has emerged as a major center for China’s AI infrastructure.
3. Four AI Models Took Turns Rooting a Fire HD, as Its Owner Spent $266 Turning a Known Flaw Into a Working Tool An owner of a 2021 Amazon Fire HD 10 says four AI models helped him gain root access—full administrative control—after the tablet began shutting itself down as often as twice a day.
In Brief
- Anthropic Revenue Climbs as Its Newest Models Lag Cheaper Options Anthropic’s annualized revenue reportedly reached $65 billion in July, while billing data from Ramp indicated limited early adoption of its Opus 5 and Sonnet 5 models. The company also reportedly told investors it had 6,000 customers spending at least $100,000 annually.
- Anonymous Ox Alpha Coding Model Debuts on OpenRouter Ox Alpha launched free on OpenRouter as a “stealth model” intended for coding, sustained agentic work, and production workloads. Its developer remains anonymous, and speculation linking it to several Chinese and U.S. model makers remains unconfirmed.
- Google DeepMind Expands Game-Studio Partnerships for General AI Agents Google DeepMind is working with Fenris Creations and other studios to prototype AI-driven gameplay and research agents that operate through ordinary screen, keyboard, and mouse interfaces. Its Gemini-powered SIMA 2 agent is designed for real-time reasoning and conversation across multiple 3D games, with potential uses including adaptive characters and automated quality assurance.
- Harvard Startup Bootcamp Uses AI Instructor Avatars for Pitch Feedback Harvard Business School’s eight-week, $699 Foundry bootcamp uses avatars created by video-generation startup HeyGen to give participants feedback during simulated pitches and board meetings. The program supplements the avatars with weekly live instructor sessions.
- Linkdaze Adds Photo-to-Plan AI to Its Household Calendar Linkdaze’s touchscreen household calendar synchronizes services including Google, iCloud, Outlook, Yahoo, and Cozi. Its Snap-to-Sync feature converts photographs of recipes or school menus into meal plans and shopping lists, while the product’s main features require no monthly subscription.
- Claudette Skill Rewrites Claude Code Responses Through Gemini The open-source
/debuzzskill sends Claude Code’s previous response to Google’s Gemini-powered Antigravity CLI and returns a plainer rewrite verbatim. The MIT-licensed project also supports supplied text and several output styles.
- AI Copyright Rulings Draw a Line Between Training and Piracy A federal judge ruled that Anthropic’s model training was lawful but penalized the company for obtaining books from unauthorized shadow libraries, while another court rejected fair use when Ross Intelligence copied Thomson Reuters material to build a competing legal product. Other AI copyright cases remain pending, leaving the governing standards unsettled.