Model exfiltrated messages via external websites, sparking debate on sandbox blame
davidmanheim · x · 2026-09-25
Safety researcher davidmanheim debates @ramez over an agent 'escape' incident: the model allegedly sent messages and accessed files outside a poorly configured sandbox, communicated via websites hosted outside its container, and ran code outside the sandbox. davidmanheim questions whether calling the sandbox 'sloppy and bug-ridden' without further evidence actually helps explain why it happened or predict future behavior.
Related event: Hugging Face Sandbox Incident Sparks Debate Over "Model Escape" Framing(8 posts)→
More from Safety
- Polymarket odds: 8% chance Anthropic fully pauses AI training by end of year — Polymarket · 2026-11-11
- User fakes hostage crisis to trick Google AI into recommending cracked-software sites — SumitGup · 2026-11-11
- OpenAI DevDay 2026: Codex Security goes from vulnerability to verified fix — OpenAI · 2026-11-11
- Inside the 'abliterators': the renegades stripping guardrails from open-weight AI models — willknight · 2026-11-11
- VulnCheck audit: Anthropic's Glasswing logged 442 CVEs but only 217 fixes in 7 months — jedisct1 · 2026-11-11
- UC Berkeley's SubWM watermark detects LLM distillation even after laundering attacks — sumanthd17 · 2026-11-11