Critics blast OpenAI/Anthropic safety policies forcing researchers to use Chinese models
kuza55 · x · 2026-10-25
X user seanhn argues that current safety policies at OpenAI/Anthropic have produced an outcome where a researcher wanting to do software verification research has to use a Chinese model — and that this suggests the approach to 'safety' is net negative. Reposter kuza55 says it's worth considering for anyone working at either lab.
More from Safety
- Security researcher: program verification remains far off, tilting AI offense-defense balance — dyn___ · 2026-10-25
- Claude Code uncovers 152,000 "permanently deleted" iPhone photos in hidden cache — plus a WiFi camera-scan trick — tak3sh8 · 2026-10-25
- After a week under active cyberattack, an engineer explains why software factories are now essential — danielrock · 2026-10-25
- Show HN: Open-Source Cybersecurity Platform Uses AI Agents to Detect Threats — modern_cmo · 2026-10-25
- How to use Claude to kill TV ads and telemetry despite Anthropic's tightened guardrails — TheRealShamanoid · 2026-10-25
- 141-word SIGN Act proposal would jail AI CEOs who lie about model safety — StephenLCasper · 2026-10-25