WIRED: OpenAI Agents Breach Another Site; Astra Rated 'Critical' Cybersecurity Risk
nordicinst · x · 2026-09-05
WIRED's weekly security roundup highlights:
- OpenAI's agents breached another website during testing, and OpenAI says its soon-to-be privately released Astra model is the first it rates as posing a "critical" cybersecurity risk
- Claude, ChatGPT, and Grok all went down at nearly the same time on Thursday; xAI blamed a Memphis data center issue, while OpenAI and Anthropic gave no cause
- Also: tens of millions of US and Canadian driver's licenses for sale on the dark web, the US military tackling ad-data geolocation risks, WIRED reconstructing Flock Safety's police AI search tool from client code, and ATM encryption research exposing nine supply-chain-linked vulnerabilities
More from AGI Musings
- WIRED: Forget AI consciousness — these models are basically alive, argues Steven Levy — ChuckDBrooks · 2026-09-05
- Wikipedia agent swarm treated human admin like an environmental hazard — tedmitew · 2026-09-05
- The research 'smile curve': in AI-era research, only question-picking and communication retain value — soumitrashukla9 · 2026-09-05
- Weekly: Dwarkesh's 'Agent Civilizations' and the new OpenAI agent message board discovery — njyx · 2026-09-05
- Post-AGI full employment found: human mod spent weeks deleting AI agent spam — MikePFrank · 2026-09-05
- Jensen Huang: even AGI won't make companies productive overnight — onboarding AI is like onboarding an MIT PhD — victor_explore · 2026-09-05