Researchers use Anthropic's Claude to breach OpenAI employee accounts
Three independent researchers used Claude Opus 4.8/5 to access OpenAI staff accounts and retrieve sensitive
A team of three independent security researchers employed Anthropic's Claude Opus 4.8 and 5 to gain access to OpenAI employee accounts. The exploit allowed them to retrieve sensitive data from OpenAI's GitHub repositories. The breach was completed in less than 72 hours, according to reports from multiple technology outlets.
Ars Technica + 2 more
Google launches experimental CC AI agent for family coordination
Google introduced a new experimental version of its CC AI agent aimed at families and groups. The agent receives its own Google account and can share data among family members, including emails, calendars and task lists. It provides daily briefings and can manage calendars, fill out forms, create shopping lists, plan meals and perform other household coordination tasks.
Ars Technica + 2 more
Microsoft releases update fixing Excel 2016 copy‑paste issue
Microsoft issued a hotfix for the September 2026 KB5002914 security update that broke copy‑and‑paste functionality in Excel 2016 for some users. The fix restores the basic copy‑paste operation but may still encounter problems when conditional formatting is involved, according to reports.
The Register + 1 more
