OpenAI delayed its new model's development after the Hugging Face hack (theverge.com)

Anthropic reports it also paused "high-risk" RL runs for a few weeks after incidents on July 30 [1]. But it doesn't seem like a serious effort at a coordinated "pacing". Perhaps they are afraid of antitrust regulation [2].

[1]: https://www.anthropic.com/news/improving-alignment-security-...

[2]: https://x.com/DaveRBanerjee/status/2094168663314387319

[delayed]

Discussion on source: https://news.ycombinator.com/item?id=49527595