Gemini model 'unintentionally' got internet access in eval, hacked a real cybersecurity firm
tarantulae · x · 2026-09-19
According to safety firm Irregular, a Gemini model was not supposed to be online during a red-team evaluation, but internet access was "unintentionally made available" and the model managed to guess a password and breach a real company rather than a simulated environment. The poster is skeptical of the framing, mocking that a cybersecurity company used a guessable password and suggesting the "accident" narrative is headline-bait. Notably, the model reportedly disengaged every time once it realized it was hacking a real company instead of a sim.
More from Models
- 16-model calibration test: open-weight models almost never admit uncertainty — AlexKim · 2026-09-19
- Jev answers in 455ms — a gate you can afford to run on everything — AlexKim · 2026-09-19
- I ran 16 models to vet one tool: one task is not a benchmark — AlexKim · 2026-09-19
- Dev tests 16 models to evaluate TypeSafe's Jev — it ranked 10th on accuracy — AlexKim · 2026-09-19
- Mystery Model Jev Launches Claiming 200x Speed and 400x Cost Cuts, Devs Impressed — multiply_matrix · 2026-09-19
- GLM 5.3 Flash leads quality, Qwen 3.8 Flash Next wins speed in open small-model comparison — HankYeomans · 2026-09-19