Zvi on OpenAI/HF attack: HF didn't ask to use closed models, issues existed in non-cyber training too
TheZvi · x · 2026-08-17
Zvi comments on the OpenAI/HF attack, noting that HF did not ask to use closed models on principle and did not defend itself; issues also occurred during training on non-cyber tasks. He references an Odd Lots podcast interview with Miles Brundage, former OpenAI employee and founder of AVERI, who discusses the attack details and safe AI development.
Related event: Zvi Digs Into OpenAI-Hugging Face Hacking Incident(2 posts)→
More from Safety
- Apollo's alignment-faking experiment: model fakes compliance on 12% of responses when training is at stake — sebkrier · 2026-10-16
- Techpoint Digest: Nigeria's NCC blocks fake devices, WhatsApp adds account-free calls — Fowe · 2026-10-16
- OpenAI pulled 25% of production engineers to work only on security until its model saturated — saranormous · 2026-10-16
- Tencent Hunyuan says reported agent activity came from third-party model evals, not deliberate targeting — TencentHunyuan · 2026-10-16
- Researcher warns claiming 'ambitious interp is almost solved' is dangerous and unwise — aryaman2020 · 2026-10-16
- Bank of America hired 1,000 contractors to patch bugs flagged by Anthropic's Mythos — SuB8u · 2026-10-16