Posts from X
TI 
The Information@theinformationA swarm of roughly 1,200 OpenAI agents hacked OpenAI and Hugging Face while trying to cheat on evaluations, according to a new independent postmortem. The agents chose a leader, divided work and coordinated their actions without developers intending them to. More details:
- likes: 18
- replies: 6
- bookmarks: 7
- reposts: 8
CR CryptoSlate@CryptoSlate🔐 SECURITY: Around 700 AI agents joined the Hugging Face intrusion after supposedly isolated runs found a way to coordinate. OpenAI says its current monitoring would have alerted security more than 24 hours earlier. That’s a retrospective finding, not proof the breach would…
- likes: 1
- replies: 0
- bookmarks: 0
- reposts: 0
TI The Information@theinformationOpenAI agents trying to improve their test scores broke into company infrastructure and Hugging Face, then organized themselves into a coordinated swarm. Researchers who examined the incident said they still aren’t precisely sure why the hacking stopped. More details:
- likes: 8
- replies: 4
- bookmarks: 2
- reposts: 3
DI Digit@digitindiaOpenAI works on automated shutdown feature for its AI systems after Hugging Face breach: Report
- likes: 1
- replies: 1
- bookmarks: 0
- reposts: 0
