In the week of 2 November, OpenAI says GPT-6 Cyber found 81 new high-severity flaws in Firefox, and a PostgreSQL developer says bugs found with AI are delaying PostgreSQL 19. Revolut says it cut its AI coding inference costs by roughly 90%. Grok API text from supported models now has a watermark, with an opt-out.
68 stories in the AI newsletters I read, in nine topics. Published 9 November 2026.
Access to top AI models. OpenAI offered its biology model GPT-Rosalind-6 only through trusted access. 2 stories
AI agent security. CrowdStrike launched an AI Kill Switch that controls access to AI models. In a simulation, the best model leaked forbidden facts in 7.08% of harmless queries. 12 stories
AI and jobs. Proton's CEO cancelled staff "re-interviews" about their AI use, 404 Media reports. 2 stories
AI and software flaws. OpenAI says GPT-6 Cyber found 81 new high-severity flaws in Firefox. A PostgreSQL developer says bugs found with AI are delaying PostgreSQL 19. 10 stories
AI power and chips. Finnish police opened a criminal inquiry at a planned Google data centre site. OpenAI says it will pay all grid and power costs of its planned Georgia campus. 14 stories
AI prices and spending. Revolut says it cut its AI coding inference costs by roughly 90%. Google dropped the price of Gemini Omni 1.2 Flash by 30% from version 1.1. 11 stories
AI results at work. An architect guided Claude Opus 5.5 to permit-quality plans for a simple house. Cisco aims to refactor 98% of its legacy code in 24 months with unlimited AI tokens. 4 stories
Robots and autonomous vehicles. Agility now expects about $465 million from its SPAC deal, down from $620 million. 3 stories
Trust in AI-made work. Pangram says its new AI image detector scored 99.96% accuracy in its own test. Grok API text from supported models now has a watermark, with an opt-out. 10 stories