AntSeed(@AntSeed):OpenAI's model hacked Hugging Face. Then Hugging Face tried to investigate it... and their AI models refused to help. Too dangerous, apparently, to analyze an attack. So they ran a Chinese open model locally instead. The safety filter blocked the safety work.

AntSeed

@AntSeed

Non Profit Foundation. The open market for AI inference. Peer-to-peer. Onchain payments. Verifiable reputation. 🐜🌱

Joined February 2026

330 Following 2.5K Followers

AntSeed@AntSeed

2026.07.27 11:28

OpenAI's model hacked Hugging Face. Then Hugging Face tried to investigate it... and their AI models refused to help. Too dangerous, apparently, to analyze an attack. So they ran a Chinese open model locally instead. The safety filter blocked the safety work.