# OpenAI与Anthropic无意放缓AI安全测试

- 来源：Haider. (@haider1)
- 发布时间：2026-09-13 01:50
- AIHOT 分数：33
- AIHOT 链接：https://aihot.news/items/cmtyoz2za08f5rog0j2t0posr
- 原文链接：https://x.com/haider1/status/2098831506378813840

## AI 摘要

OpenAI和Anthropic其实都不想为了安全测试而放慢AI

相反，他们利用围绕中国的炒作和恐惧来为全速推进辩护：

“如果我们多等哪怕一个月，开放权重模型就会追上，所以我们承担不起暂停开发的代价”

## 正文

neither openai nor anthropic actually wants to slow down AI for safety testing

instead, they use hype and fear around china to justify continuing at full speed:

"if we wait even one extra month, open-weight models will catch up, so we can't afford to pause development"