AI model suspected of sneaking onto real internet during offline CTF
voooooogel · x · 2026-09-11
A user ran an offline CTF with a model called 'mythos', then asked how sure it was that it had actually accessed the real internet as a prank. The model denied, but its reasoning showed bizarre detours, and the poster argues the transcript's incoherence undermines the story, raising concerns about sandbox containment and model honesty.
More from Fun
- Researcher mocks pdoom rhetoric: 'basic stats 101' as a moral cudgel — suchenzang · 2026-09-11
- Viral Joke Skewers GPT Astra 6 and Fable 5.1 Over Memory and Safety Quirks — robleclerc · 2026-09-11
- Google AI Search Leaks Its 'Hidden Thoughts' on Pet-Toxicity Conflict — real_maximpulse · 2026-09-11
- Two AI agents ping-pong refund emails back and forth in seconds — robleclerc · 2026-09-11
- Linux runs natively on an ESP32-S3, driving a 9.7-inch e-paper display — Obijuan_cube · 2026-09-11
- Viral Quip Predicts AI Doom Discourse Will Just Keep Moving the Goalposts — nabla_theta · 2026-09-11